Optimizing sustainability of infrastructure projects through the integration of building information modeling and envision rating system at the design stage

Sustainable Cities and Society 84 (2022) 104013 Contents lists available at ScienceDirect Sustainable Cities and Society journal homepage: www.elsevier.com/locate/scs Optimizing sustainability of infrastructure projects through the integration of building information modeling and envision rating system at the design stage Avin Laali a, Seyed Hossein Hosseini Nourzad a, *, Vahid Faghihi b a b Department of Construction and Project Management, School of Architecture, University of Tehran, Tehran, Iran Construction Science Department, School of Architecture, Prairie View A&M University, Prairie View, TX, United States A R T I C L E I N F O A B S T R A C T Keywords: Infrastructure Building information modeling Sustainable design Optimization Envision Sustainability evaluation systems With infrastructure services demand rising, a significant contributor to sustainable development goals, infrastructure systems have to evolve and adapt effectively, efficiently, and sustainably. The broad set of decision variables coupled with the trade-off among the economic, social, and environmental aspects complicate sustainability assessment and optimization during the design stage. Thus, this paper presents a Building Information Modeling (BIM)-based automated framework for real-time evaluation and optimization of sustainability in infrastructure projects. The proposed framework benefits from infrastructure sustainability rating systems that provide a systemic and balanced set of indicators for careful consideration of sustainability. On the other hand, to automate the process, BIM, with its 3D shared environment, is used as an integrated platform for dynamic sustainability analysis. The proposed framework integrates both novelties in real-time during the early design stages, which helps designers to select the most sustainable alternative. A prototype and a hypothetical case study are conducted to validate the framework’s applicability through Infraworks 360 as the operating BIM platform and Envision as the baseline rating system. The results confirm that adopting the suggested method simplifies the inclusion of sustainability into design decisions while facilitating documentation of compliance with credit assessment for both project teams and verification agencies. 1. Introduction Despite the environmental consequences, humankind seems unwilling to give up the services provided by the facilities they have built. As a result, they seek refuge in the most known definition of the concept of “Sustainable Development,” which has attracted significant attention in recent years as development that serves current needs without jeopardizing future generations’ ability to meet their own (WCED, 1987). This statement implies that we want to develop such amenities while minimizing or even reversing their adverse impacts. In addition, the 17 Sustainable Development Goals (SDGs) issued by the United Nations have challenged every industry worldwide to realign its strategies and efficiently carry out sustainable objectives (Marzouk & Othman, 2020; UN, 2015). As a great contributor to global warming, the construction industry is also under extreme pressure to stop such damages and go beyond measures to repair such losses (Tryggestad, 2013). Infrastructure projects as a central and impactful part of this industry contribute directly to SDGs 9 and 11 (Clevenger, Ozbek & Simpson, 2013; UN, 2015). Therefore, sustainable development and planning of infrastructures become a must as also issued by. However, infrastructure design is generally complicated because of the various factors and considerations involved in their design, which define significant limitations and strictly condition their design (Bongiorno, Bosurgi, Carbone, Pellegrino & Sollazzo, 2019). Additionally, sustainability is also commonly thought of and analyzed in terms of the balance between different economic, social, and environmental dimensions, which is not readily apparent to decision-makers and stakeholders during the early design process (IDB, 2018; Shahtaheri, Flint & De LA Garza, 2018). Thus, finding an optimum solution becomes even more of a complex task. These delicate tasks necessitate the provision of timely, dependable, and practical solutions that can meet multiple, potentially contradictory aspects of sustainability. Therefore, many efforts have been made to develop novel tools and algorithms to support the designers in this intricate task (Rees, 2010). Building Information * Corresponding author. E-mail addresses: avin.laali@ut.ac.ir (A. Laali), hnourzad@ut.ac.ir (S.H.H. Nourzad), vafaghihi@pvamu.edu (V. Faghihi). https://doi.org/10.1016/j.scs.2022.104013 Received 9 March 2022; Received in revised form 20 May 2022; Accepted 18 June 2022 Available online 19 June 2022 2210-6707/© 2022 Elsevier Ltd. All rights reserved. A. Laali et al. Sustainable Cities and Society 84 (2022) 104013 2. Literature review Modeling (BIM) tools are among these proposed solutions and novelties. BIM can be characterized as an integration of processes and technologies that run throughout an asset’s development phases to facilitate the generation, storing, managing, and sharing of physical & functional data. In short, BIM is a digital representation of the facility rich with data and parametric rules that allow intelligent modifications (AGC, 2005). BIM stores data through artificial intelligence (AI) and by using specialized “smart objects” that depict the realistic element in the 3D scene. Each smart-object consists of features that guarantee a full comprehension of the object’s qualities, role, interactions, and influence on the external context, enabling various consistent and integrated elements analyses (Chen, Lok, & Jeng, 2016). Therefore, when changes are made, smart objects are automatically updated (e.g., any alterations to the horizontal alignment, vertical profile, or road cross-section of a road have an instantaneous impact on the other representations) (Lorek, 2018). In general, this innovation can considerably minimize errors and inaccuracies due to the human mind’s cognitive limitations (Hergunsel, 2011). Sustainability rating systems can be regarded as the product of the same objectives, recognized as frameworks that provide comprehensive coverage of all sustainability aspects and their scientific interrelationships, allow setting of sustainable priorities and goals, and enable effective stakeholder communication (Ando et al., 2005; Cole, 2003; Liu, Van Nederveen, Wu, & Hertogh, 2018). At the same time, since the quantitative analysis of data is critical for digital sustainability assessments, the quantitative approach to viewing the sustainable development of these frameworks can serve as the foundation for such efforts. However, their utilization can be time-consuming due to the interactive data computation that seeks the best fitting design that achieves the desired degree of sustainability (Carvalho, Bragança & Mateus, 2019). Despite the infrastructure sector’s proven potential and greater impact, the building industry is well ahead in developing, applying, and integrating both novelties (Liu, Van Nederveen, Wu, & Hertogh, 2018; Mcvoy, Nelson, Krekeler, Kolb & Gritsavage, 2010). For example, Biswas, Wang & Sung-Hsien (2008) created a program that uses BIM technology to evaluate the environmental effects of design decisions. More recently, Carvalho et al. (2019) examined how BIM might aid in the optimization of building sustainability assessment methodologies, concentrating on the SBToolPT-H (Sustainable Building Tool (SBTool) in the context of Portuguese residential buildings). The findings of the SBToolPT-H reveal that BIM may be used to examine 24 of the 25 criteria directly or indirectly. Their research comprised one of the first attempts to integrate BIM into green building grading and certification. Some studies validate the same theory in the infrastructure sector but rarely present practical answers while struggling with finding the proper platforms and interoperability issues (Liu, Van Nederveen, Wu, & Hertogh, 2018). Hence, this paper’s novelty addresses the lack of awareness of the potentialities of BIM for the effective and automated use of sustainability rating systems as benchmarks to develop a sustainability optimization method. Hence, this study aims to bridge the mentioned gaps by proposing a practical solution and presenting an automated BIM-based framework for sustainability optimization at the design stage of infrastructure projects. In this way, infrastructure designers can instantly be informed of a potential design alternative’s sustainability impact. They are also presented with optimized sustainability alternatives based on the baseline design and sustainability evaluation system requirements. The framework’s applicability will be tested by developing a prototype that uses Infraworks 360 (Chappell, 2015) as the BIM platform and Envision, version 3 (ISI, 2018b, 2021) for a sustainability evaluation system for road alignment projects. This prototype employs a genetic algorithm (GA) as the operating optimization algorithm to support decision-makers with sustainably optimized and modifiable BIM-based solutions. 2.1. Sustainability and infrastructure sustainability assessment tools Although holistic definitions of sustainability have generated sustainable construction industry policies, they are simply not detailed enough. As a result, the demand for establishing and evaluating environmental performance indicators led to the development of several sustainable “building” rating systems (Cole, 2003; Diaz-Sarachaga, Jato-Espino, Alsulami & Castro-Fresno, 2016). Afterward, broadening these assessment tools’ scope to quantify the entire built environment’s impact on sustainability pillars resulted in the development of sustainability rating tools for infrastructure projects (Ferrer, Thomé & Scavarda, 2018; Mcvoy et al., 2010). Although various sustainability assessment tools and rating systems are available, few have the integral approach reflecting various aspects of sustainable development that enable assessing a wide range of infrastructure projects (Diaz-Sarachaga et al., 2016). Envision in the USA (ISI, 2018b, 2021) along with CEEQUAL in the UK (BreGlobal, 2021), and the Infrastructure Sustainability rating scheme in Australia (ISCA, 2021) is among those few that have the integral approach that reflect various aspects of sustainable development of a wide range of infrastructure projects regardless of their type or size. Due to the relatively recent introduction of these tools, there are limited project-level studies. Studies have typically concentrated on the overall benefits and drawbacks of rating tools (Ainger & Fenner, 2014; French, 2012; Hurley, 2009) and evaluations of individual rating tools (Willetts, Burdon, Glass & Frost, 2010). Griffiths, Boyle and Henning (2017) observed that the tools address comparable impact areas, but the content, strategy, and opportunities to improve sustainability performance vary. The observed that CEEQUAL rewarded incremental actions more than the other tools, which is likely to suit project teams new with sustainability principles and techniques. Moreover, Envision was noted as a valuable reference enabling early project planning and assessment of larger sustainability and community concerns. It also provides guidance to infrastructure owners who want to push the envelope in terms of restorative efforts and long-term planning. IS, like Envision, placed a premium on stakeholder and community participation in sustainability actions and decisions, which had previously been identified as critical to achieving sustainable outcomes. Envision, established in 2012 by the Harvard Graduate School of Design’s Zofnass Program for Sustainable Infrastructure and the Institute for Sustainable Infrastructure (ISI), is intended to evaluate and score the overall contribution of infrastructure projects to sustainability based on the TBL (ISI, 2018a). This comprehensive approach to infrastructure development seeks to assess projects in terms of their value to communities, efficient use of funds, and contribution to long-term sustainability. Envision also considers all aspects of the lifecycle, allowing for better-informed decisions in all stages of planning to deconstruction or decommissioning (Gaughan, 2012; Saville, Miller & Brumbelow, 2016). Although Envision is a self-assessment tool, ISI provides an optional third-party verification mandatory for awards. The rating is based on 64 criteria, known as credits, organized into five categories: Quality of Life, Leadership, Resource Allocation, and a unique category of Natural World and Climate and Risk (Infrastructure & Infrastructure, 2018; ISI, 2018a, 2015). Credits are grouped to assist users in managing the complicated trade-offs and synergies between them. Each credit is assigned points weighted in line with its estimated contribution to sustainability. Each of them is rated according to the following achievement levels, from lowest to highest: Improved, Enhanced, Superior, Conserving, and Restorative. However, there are not five degrees of achievement for every credit as the level degrees are established by the credit’s type and the capacity to distinguish between 2 A. Laali et al. Sustainable Cities and Society 84 (2022) 104013 levels. Credits that do not apply to the specific project are marked as (N/A1), thus removing them from consideration. received much attention, sustainability evaluation and improvement through BIM are yet to receive practical solutions. In a study performed by Liu, Van Nederveen, Wu, & Hertogh (2018), a conceptual framework based on interviewees’ assessments combined with a critical analysis of technical requirements guiding the integration is presented, allowing for what-if scenarios to better support incorporating sustainability into design decisions. In a literature review of 471 scholarly bibliographies performed by Liu, LU, Shen and Peh (2020), one of the four main identified research gaps was the lack of a comprehensive overall building performance framework to support BIM. Of the few and most recent studies addressing both of the considerations, van Eldik et al. (2020) used BIM as a platform for automation of the environmental impact assessment in the early design stages of infrastructure projects to provide designers with accurate results of the environmental impact of all objects associated with the projects’ design. Smart cities and growing cities need BIM input to help ensure a stronger adherence to sustainability in our fragile and ecologically unstable environment (Marzouk & Othman, 2020). Engaging sustainability considerations in construction projects’ processes can contribute to more complexity (Rahmani Asl, Zarrinmehr, Bergin & Yan, 2015). Besides, there is growing attention to the use of BIM as an all-encompassing source of data to optimize computationally complex problems that require considerable amounts of data for more accurate analysis. Marzouk and Othman (2020) integrated BIM and the geographical information system (GIS) to plan smart cities, starting with the utility infrastructure for growing cities and those cities still in the formation stage. They considered wastewater management and freshwater protection, as well as electrical needs. After dividing the city into different plots, they analyzed land use, building features, and other information to help them build the smart city development strategies. 2.2. Design stage Undoubtedly, measures must be taken at the early design stages to effectively evaluate and enhance an asset’s nearly permanent and widespread sustainability impacts (Azhar, Carlton, Olsen & Ahmad, 2011). Many studies support this, demonstrating the point of action that suggests the most potential for sustainability improvements are the early decision-making stages of an asset (Basbagill, Flager, Lepech & Fischer, 2013; Bueno & Fabricio, 2018). Shahtaheri et al. (2018) proposed a multi-criteria preference assessment framework for decision-makers to make more sustainably informed decisions regarding early design alternatives for commercial buildings. Sabatino, Frangopol and Dong (2015) presented a sustainability-based maintenance optimization decision support for highway bridges on the transportation front. However, sustainability solutions cannot be formed when there is a lack of adequate data regarding the sustainability impacts of different strategies and design alternatives (Shahtaheri et al., 2018). Especially in infrastructure projects with a more sizeable interaction with their surrounding environment, creating a realistic, responsive 3D model of that environment becomes ever more obligatory (Bongiorno et al., 2019). Therefore, BIM as a modeling platform that captures physical and functional interactions for sustainability optimization becomes evident. While 3D modeling has become more widely used in building design and engineering, 2D drawing-based procedures are still commonly used in infrastructure (Chong, Lopez, Wang, Wang & Zhao, 2016; Kim, Kim, Ok & Yang, 2015). Bradley, Li, Lark and Dunn (2016) observed that despite the numerous studies dedicated to BIM-based research at the design stage of building projects (Hollberg, Genova & Habert, 2020; Santos, Costa, Silvestre & PYL, 2020; Wong & Zhou, 2015), few solutions had been presented for infrastructure projects (Minagawa & Kusayanagi, 2015). Nevertheless, recent research aims to remove the barriers to BIM adoption in the infrastructure domain (Chan, Olawumi & Ho, 2019; Hartmann, Van Meerveld, Vossebeld & Adriaanse, 2012; Ji, Borrmann, Beetz & Obergrießer, 2013). By referring to best practices in the building domain, the aim is to examine and utilize BIM’s potential for improving the environmental impacts of designs (Liu, Van Nederveen, Wu, & Hertogh, 2018; van Eldik, Vahdatikhaki, Dos Santos, Visser & Doree, 2020). Several are crucial for improving sustainability evaluation and optimization among these potentialities, as Bongiorno et al. (2019) mentioned. These include but are not limited to the “smart objects” in the BIM environment that reflect the project’s key elements in 3D, representation of the individual pieces and their interactions, and an extensive relational database of all of the relevant linked information that can influence and condition the many project phases. 2.4. BIM and optimization While the assessment and comparison of sustainability impacts of single alternatives can appear to offer all the right solutions, generating a globally optimal model through such methods can become highly complex. Optimization methods can achieve high-performance alternatives while overcoming the difficulty of balancing conflicting and potentially complex objectives of infrastructure projects, especially under TBL objectives, for maximum performance achievement (Nguyen, Reiter & Rigo, 2014; Rahmani Asl et al., 2015; Thangaraj, Pant, Abraham & Bouvry, 2011). Despite the proven benefits of BIM as a consistent and parametric platform, not many studies that propose sustainability optimization frameworks at the design stage of infrastructure projects utilize it as an accurate source of necessary project information that also provides dynamic access to decision variables (Bongiorno et al., 2019; Rahmani Asl et al., 2015; van Eldik et al., 2020). According to Geyer (2009), applying optimization methods at the design stage of construction projects has a twofold benefit: it may lead to better-performing solutions and a greater understanding of the design space or the range of viable options. The design of infrastructure projects can be complex due to the many involved factors and objectives coupled with sustainability issues (Shahtaheri et al., 2018). Optimization algorithms are a powerful method for supporting this process by comparing thousands of solutions with low computation efforts (Bongiorno et al., 2019; Geyer, 2009). While sustainability applications of BIM for buildings and especially optimizing their sustainability issues have become more sophisticated over the past 20 years, knowledge and practical solutions in infrastructure projects are scarce (Liu, Van Nederveen, Wu, & Hertogh, 2018). Among studies that prove the beneficial synergies of combining optimization method and BIM for sustainable development of buildings is a study performed by Liu et al. (2015). This study used Ecotect Analysis by Autodesk Inc. (Yan et al., 2009) for a Particle Swarm Optimization (PSO)-based optimization of life cycle costs (LCC) and life cycle carbon emissions (LCCE) of building designs. Also, Rahmani Asl et al. 2.3. BIM and infrastructure sustainability Infrastructures are supporting systems to sustain human civilization and activities. The overall scope of infrastructure development includes transportation systems, which are viewed as the foundation of a prosperous economy and social development (Costin, Adibfar, Hu & Chen, 2018). Notably, efficient and innovative technologies are needed to support the development of new infrastructure services and, often, the replacement of aging transportation structures (Adshead, Thacker, Fuldauer & Hall, 2019). Among these technologies, BIM stands out with its 3D modeling technology, a design process that has changed how the construction industry operates. The building information model is interpreted as the product, while the building information modeling is described as the process (Hooper & Ekholm, 2012). Although BIM employment in infrastructure projects has recently 1 Not Applicable. 3 A. Laali et al. Sustainable Cities and Society 84 (2022) 104013 (2015) presented a building performance optimization based on performance simulation on top of Revit, a widely used BIM platform. This combination could generate and assess building model designs and search for the best sustainably accomplished options to aid designers in performing sustainability performance design analysis. As stated, very few studies addressed the potentialities of BIM as an enabler to access full range of information for sustainability performance optimization of infrastructure projects. As such, Bongiorno et al. (2019) investigated the possibility and advantages of such integration in the case of 3D highway alignment optimization. A study performed by Bongiorno et al. (2019) is among the few studies that address the possible advantages of integrating optimization of highway alignments with BIM serving as a comprehensive source of data and enabler of model visualization and modification. criteria, the sustainability rating system (Envision) determines the rules and formulas for assessing the infrastructure’s sustainability performance, hence, the sustainability constraints and fitness function of the optimization algorithm. In this study, Envision system was selected as the baseline rating system. The main reasons for choosing Envision in this study are as follows: 3. Method Therefore, in this study, one of the mainstream sustainability assessment tools, Envision, was selected as the baseline evaluation tool for the comprehensiveness and scalability of the framework. Additionally, the I-BIM relational database can serve as a comprehensive data source for the optimization algorithm. This is done through the smart-object information included in the BIM relational database. Smart objects are not only a realistic representation of any element in a 3D scenario but rather a collection of features that ensure an in-depth representation of each object’s characteristics, roles, interactions, and influence on the external context. This study will use Autodesk Infraworks 360 as the BIM platform for implementing the proposed frameworks as the most compatible BIM tool concerning its capabilities of modeling the selected infrastructure type (roads) and its decision variables through road alignment smart objects in a realistic manner. For instance, for any road alignment smart object, a design speed can be determined manually and automatically that specifies the minimum and maximum radius, spiral length, and tangent length based on predefined road design standards (e.g., AASHTO standard). Infraworks 360 was used as the authoring platform to evaluate the integrated BIM-based framework’s applicability and develop the prototype in JavaScript format using the scripting console feature to access and manipulate data through the API. This platform can also incorporate project and common data through modeling or importing features. Lastly, once the objective function, which maximizes achievable points from Envision, and the design and environmental constraints are set, the optimization algorithm could be run using the scripting console of the BIM tool to load the developed optimization algorithm’s coding. Outputs are then produced, being the optimal solution modeled and a) This rating system can ensure and evaluate the sustainability of all types of infrastructure projects of different sizes. b) It can also be flexible since “the final Envision score is presented as a percentage of the total applicable points.” (Envision, 2020). c) The final reason for selecting Envision is the inter-relation of its credits which enlightens the effects of gaining points from one credit based on other credits. This research was carried out to develop a BIM-based framework for the sustainability optimization of infrastructures. A prototype was developed and tested on a hypothetical case study to validate the framework’s applicability. 3.1. Framework for BIM-based sustainability optimization The framework’s broad applicability is provided through its modular design, allowing methods, tools, and types of each of the four components to be altered: sustainability rating tool, infrastructure project, optimization algorithm, and the BIM platform, as shown in Fig. 1. Fig. 1 also represents the inter-relationship among these primary elements and their transactions. The infrastructure type indicates the applicable criteria from the baseline rating system and design constraints and parameters for the optimization algorithm. In this study, road alignments were selected to implement the framework. Transportation infrastructures are a vital part of any civilization’s economy, safety, and well-being. There is a clear need to advance the existing and future transportation infrastructures to adjust and go beyond the growing concerns revolving around sustainability issues (Costin et al., 2018). Additionally, this choice was made due to the congested presence of this infrastructure type, with over four million miles of road network in the United States and an annual average increase of 0.4 percent, making it one of the most prevalent and vital forms of infrastructure (USDOT, 2018). In addition to providing a comprehensive set of sustainability Fig. 1. The basic components of the framework. 4 A. Laali et al. Sustainable Cities and Society 84 (2022) 104013 displayed in the BIM platform as a road alignment 3D smart object and its detailed points achieved from the sustainability evaluation rating system. As mentioned, the BIM platform provides the input for calculating the required value by the rating system to create an optimal solution. The developed prototype is based on a road alignment infrastructure project design in this study. Among the various mathematical models and optimization methods, genetic algorithm (GA) approaches and the swarm intelligence (SI) methods are both iterative and evolutionary techniques. They provide the benefits of efficiency and easy model formulation and have proven reliable. The first study for a GA-based search model for road alignment optimization was performed by Jong (1998). That research considered alignments as a series of 3D points of intersection (PIs) distributed in equally distanced vertical cutting planes. Many other studies followed to perfect the model’s efficiency and results (Jong et al., 2000; Kang et al., 2007; Kang et al., 2009; Kim et al., 2007). The model formulation in this study is inspired by AlHadad’s work (Al-Hadad, 2011). This study considers any road alignment as a series of station points that are often spaced evenly and define the final alignment. Unlike former methods that determine the placement of these points secondary to the alignment form when completing the final details, we looked at the series of station points that make up the alignment as a primary to the alignment form. proposed road alignment optimization models as minimization of their cost components were studied to determine the Envision’s credits equivalent to the cost components for validating their applicability to road alignments as presented in Figs. 3 and 4. As some decision variables are not evaluated under credits from Resource Allocation and Climate and Resilience categories, the horizontal curvature, vertical curvature, and grade violation cost components do not have an immediate link to these credits (Fig. 4). These cost components include but are not limited to locationdependent, length-dependent, and earthwork costs. In this approach, the best solution costs the least as the optimization model is objective function is formulated to minimize the overall cost of the potential solution (Kang, Jha & Schonfeld, 2012). The design constraints in these algorithms are imposed as penalizing costs, meaning that violating these design constraints leads to cost increment of the generated solutions and, therefore, decreases their fitness (Al-Hadad & Mawdesley, 2010). The applicability of these credits to the project type was based on the indicated cost components and sustainability design items of highway projects proposed by Tsai and Chang (2012). Since quantitative criteria are more compatible for assessment and optimization automation to eliminate the possibility of cognitive bias, the sustainability assessment tool was studied to select quantitative criteria in the second step. Table 1 presents an example of analyzing two credits from Envision to distinguish quantitively measurable credits from quantitative credits and identify the necessary data for the evaluation and the BIM method that may be utilized to acquire the relevant data. Former studies that proposed road alignment optimization models mainly consider the objective function as a minimization function of conflicting cost components (i.e., a decrease in one cost component may increase another). Hence, the lower a solution’s total cost, the fitter it is. The definition of the optimization problem in this paper is based on conflicting criteria as well; however, instead of cost components, sustainability credits form the objective function. (Al-Hadad & Mawdesley, 2010; Jha & Schonfeld, 2000; Tsai & Chang, 2012). Thus, the objective function is a maximization of the achievable scores from Envision credits, which are inherently conflicting. Hence, the higher the score, the fitter the solution. The selected sustainability assessment criteria, defined and decoded as mathematical rules and indicators, are then translated into sustainability score achievement scenarios based on the infrastructure’s design parameters and decision variables. These evaluation scenarios are integrated with the BIM tool to form an automated sustainability evaluation functionality that can be applied to any proposed infrastructure BIM-based model and return the achieved sustainability score, as shown in Fig. 5. Lastly, the optimization problem formulation is performed based on the BIM tool’s data formats and the integrated sustainability assessment algorithm, translating into the objective function. 3.2. Model implementation The proposed framework is implemented in two main phases: (1) developing a BIM-based sustainability assessment algorithm and (2) developing a BIM-based sustainability optimization algorithm. The development of the sustainability assessment algorithm is required to evaluate the produced solutions (i.e., chromosomes) in the optimization algorithm. The procedure for implementing the proposed framework is depicted in detail in Fig. 2. 3.3. Case study A potential highway alignment was considered to test the designed prototype’s efficiency and applicability in sustainability assessment and optimization of the alignments and the proposed method’s accuracy. The study area is located between Alachua and Bradford counties in the southeastern region of Florida, USA. The start- and end-points of the potential highway connect CR-18 and CR-1471 in Florida. Although the potential road alignment is hypothetical, real-world geographic data is used for this case study. The study site area is about 782 miles as the model builder feature of the Infraworks platform allows a maximum of such area for creating BIM models of potential projects. The Euclidean distance between the start- and end-points is about 8 miles. The reason for this choice of the case study area was threefold. (1) Although Envision can be used worldwide, it is a US-based sustainability rating system used extensively in the USA and Canada. Therefore, the study area was considered in the USA for this study’s purposes and the selected credits’ applicability. (2) This study’s scope only examines a 2D highway model optimization. Therefore, the suggested alternatives were given fewer variations in terms of earthwork; to be pragmatic, the state of Florida was selected as it has low height variations. (3) Access to accurate data concerning the geographical locations of wetlands and conservation areas from this area in the formats of shapefiles was possible. 3.5. The BIM-based sustainability optimization algorithm The development of the optimization algorithm in this study can be summarized in five main steps: 1 Formation of chromosome structure appropriate to the problem; 2 Determination of the evaluation conditions (objective function); 3 Creation of a random initial population of chromosomes within the boundary of the study area (initial answer); 4 Selection of an appropriate mechanism for producing offspring (selection, crossover, mutation); 5 Termination criteria must also be set to stop the algorithm when those conditions are met. 3.4. The BIM-based sustainability assessment algorithm The first phase in achieving the study’s objective is creating a sustainability assessment algorithm based on the selected sustainability rating system. This module will provide the means for calculating the fitness of the produced chromosomes in the genetic algorithm. As the first step, applicable credits from Envision to the selected infrastructure should be identified. For that purpose, former studies that have 3.5.1. Chromosome representation The first step of formulating a GA optimization is setting the proper chromosome configuration structure, which will be based on the 5 A. Laali et al. Sustainable Cities and Society 84 (2022) 104013 Fig. 2. Model formulation and framework implementation procedures and the expected outputs. infrastructure evaluation parameters and the data format of the optimization problem’s search area. This paper addresses the assessment and optimization of horizontal road alignments; as such, the model formulation is based upon the fact that any road alignment created with whatever method can be defined by the coordinates of its PIs. Therefore, the suggested optimization model considers the X and Y coordinates of the PIs as decision variables for horizontal alignment optimization, and the arrangement of its PI series can determine the placement of the highway alignment. This model uses the GA to find the best path axis candidate. In this model, the road alignments are represented by chromosomes combined of genes defined by the intersection points’ 2D coordinates (x, y). The search area is a rectangle determined by the startand end-points of the proposed baseline alternative. The data format of the study area is a key point in designing the algorithm’s formulation. The development of both algorithms must be structured to suit the accessible data format from the BIM platform for each element of the study area and decision variables for the intended credit assessment. The framework’s primary foundation is based on the infrastructure’s BIM model due to the parametric modeling advantages of BIM that facilitate the setup of an asset’s decision variables. The smart objects in this study (i.e., roads and coverage areas) are represented as GeoJSON geometry types developed by the International Organization for Standardization (ISO) and the Open Geospatial Consortium (OGC) (Schmid, Galicz & Reinhardt, 2015). These are the primary data format used in the BIM-based model, meaning that their connecting points’ coordinates define each object, as represented in Fig. 6. Thus, the model consists of objects defined by their coordinates and additional properties and attributes, as shown in Fig. 6. The BIM-based road smart objects are LineStrings, a sequence of 2D points considered as “PIs2” connected by line segments (with fixed slopes). Hence, the design of the road alignments is translated into the proper placement of their X, Y, and Z coordinates. Since the final proposed alignment inherits the attributes of the input “smart object” (e.g., road width, assembly components, materials, number of lanes, design speed, and slope), it is important to adjust the input smart objects’ attributes accordingly. In this study, “planning roads” were selected as the input “smart object,” representing a 2-lane road with 40 ft width. The planning road design speed, a customizable feature, is set as 70 mph, which determines the overall geometry of the road (e.g., degree of horizontal and vertical curvatures). Other attributes like the road’s width determine the overlapping area with environmentally protected areas, hence, entirely changing the assessment of a potential solution. Therefore, changing the input smart-object type and other adjustable properties of the BIM-based road can affect the potential output optimum solutions. 3.5.2. Objective function The second step is defining the objective function and optimization constraints. Since this study aims to present a framework for sustainability optimization of infrastructure projects by integrating BIM and sustainability assessment systems, the objective function can be defined 2 6 Points of Intersection. A. Laali et al. Sustainable Cities and Society 84 (2022) 104013 Fig. 3. Relevant credits from Envision system to cost components of road alignments (i.e., Natural World and Quality of Life categories). as maximizing the total achievable points from the rating system. Thus, the objective function can be defined as Eq. (1): Max Credits = nIP ∑ ai Crediti natural buffers from wetlands, shorelines, and waterbodies. In other words, the farther the project gets from these sensitive areas, the higher the project score. The project team is required to provide evidence and documentation to prove intentional avoidance of these areas. These rules also apply for NW1.3, which considers the percentage of farmland avoided or preserved during infrastructure development as the metric for evaluation. The highest possible level of achievement from these credits does not fall within the scope of this study. The detailed scenarios with their scores for determining the level of achievement are presented in Table 2, which translate into Eqs. (2) and (3). In Infraworks 360, these environmentally sensitive areas are modeled as coverage areas, a closed polygonal chain consisting of a finite number of straight-line segments, shown in Fig. 7. (1) i=1 As stated, a horizontal road alignment optimization often focuses on potentially opposing cost factors that deal with the location and length of the alignment (Al-Hadad & Mawdesley, 2010; Kang et al., 2012; Maji & Jha, 2009). Therefore, four of Envision’s quantifiable credits were selected that are, on the one hand, related to the horizontal alignment decision variables and are also potentially conflicting decision objectives. These credits include NW1.2, NW1.3, CR1.1, and CR1.2. The first two credits are location-dependent sustainability criteria (NW1.2 and NW1.3). Suppose the generated alignments avoid affecting these environmentally sensitive areas. In that case, they have a better sustainability performance. Up to a certain point defined in the sustainability assessment system, the further they get from these environmentally sensitive lands, the higher their fitness score. The latter two credits (CR1.1 and CR1.2) are length-dependent sustainability criteria related to road alignment design. Thus, the shorter the length of the generated alignments, the more their fitness improves. These selected criteria are potentially conflicting credits in the design of road alignments since avoiding the environmentally sensitive areas might add to the length of the alignment and vice versa. The rules defined in Envision regarding the evaluation of the NW1.2 criterion require the project to provide specific amounts of vegetated or 2, n ∑ k=1 CrNW1.2 = { n ∑ 0.9Ai , minD ≥ 50 i=1 { } 6, 50 < minD ≤ 100 , D = dj 12, 100 < minD ≤ 150 n n ∑ ∑ 16, 0.9Ai , minD ≥ 150 lk × W < k=1 7 lk × W < i=1 n j (2) Sustainable Cities and Society 84 (2022) 104013 A. Laali et al. Fig. 4. Relevant credits from Envision to cost components of road alignments (i.e., Resource Allocation and Climate and Resilience categories). Table 1 Analyzed credits from Envision. metrics intent extent BIM use for evaluation Natural World Siting NW1.1 Preserve Sites of High Ecological Value Avoidance of high ecological value sites and establishment of protective buffer zones. Avoid placing the project and temporary works on a site that has been identified as being of high ecological value. Habitat Expansion Climate and Resilience Emissions CR1.1 Reduce Net Embodied Carbon Percentage of reduction in net embodied carbon of materials. Reduce the impacts of material extraction, refinement/ manufacture, and transport over the project life. 50% In BIM, tools such as Infraworks avoidance areas can be set in the way of the development of the projects. The ability to import data from GIS enables designers to import maps of sites with high ecological value and define needed buffers to avoid these areas. The modules to calculate the possible existing overlap enable partial assessment of this credit. Creating a database of different types of high ecological value sites and the associated buffers and protection plans is recommended. Quantity take-off features enable designers to evaluate the materials needed for different design alternatives, and from there, it is a straightforward process of calculating baseline and alternative comparison. 2, CrNW1.3 = { 8, n ∑ lk × W < n ∑ k=1 i=1 n ∑ n ∑ lk × W < k=1 12, n ∑ k=1 that reduces used materials and, ultimately, reduces overall project impact. CR1.2 credit intends to reduce the project’s share of contribution to climate change by reducing greenhouse gas emissions during the project’s operation stage. When it comes to road alignments, this can be viewed as a reduction in road length because a longer road means an increase in vehicle travel distance and thus higher greenhouse gas (GHG) emission levels. Table 3 presents the detailed performance evaluation scenarios. Eqs. (4) and (5) demonstrate how the percentage of length reduction translates into the achievable credit scores. 0.1An 0.5An (3) i=1 lk × W < n ∑ An i=1 Where lk is the length of the alignment overlapping with the environmentally sensitive area (k), W is the width of the area covered by the alignment, Ai is the area of the environmentally sensitive area (i), dj is the distance of the (jth) segment of the alignment. The CR1.1 credit intends to reduce material extraction, refinement/ manufacture, and transport impacts over the project’s life cycle. The calculations for determining the achievement level in this credit are based on the percentage of reduction in the net embodied carbon of materials over the project’s life compared to the baseline. This can also be viewed as designing projects to use fewer materials. In the case of road alignments, this can be interpreted as a reduction in road length 5, ln ≤ 0.95 lb CrCR1.1 = { 10, 0.95 lb < ln ≤ 0.85 lb 15, 0.85 lb < ln ≤ 0.8 lb (4) 8, ln ≤ 0.9 lb CrCR1.2 = { 13, 0.9 lb < ln ≤ 0.75 lb 18, 0.75 lb < ln ≤ 0.5 lb (5) Where ln is the length of potential alternative alignment, lb is the length of the baseline alignment. 8 A. Laali et al. Sustainable Cities and Society 84 (2022) 104013 Fig. 5. The BIM-based Sustainability Assessment Module Output (total fitness (score) is the sum of all the achieved scores from four Envision credits represented as na12, na13, cr11, and cr12). Fig. 6. A wired representation of the selected search area, representing coordinated data style of objects. Additionally, an alternate approach was devised for constructing performance evaluation scenarios for C.R.1.1 and C.R.1.2. The length reduction scenarios (C.R.1.1 and C.R.1.2) were converted into length increment evaluation scenarios, i.e., length increment concerning the Euclidean distance between the baseline design’s start- and end-points percentage. This conversion was done due to two reasons: 1. Suppose a proper baseline alignment is not provided for the assessment to occur, 2. This scenario design has a better restricting effect since the comparison is made to the shortest possible alignment. The alternative scenarios for length reduction are described in Table 4. Eqs. (6) and (7) are 9 A. Laali et al. Sustainable Cities and Society 84 (2022) 104013 Table 2 Detailed level of achievement scenarios of credits NW1.2 and NW1.3. Natural World Siting NW1.2 Provide Wetland & Surface Water Buffers NW1.3 Preserve Prime Farmland Improved Enhanced Superior Conserving Restorative 2 (avoids 90% of wetland with a 50 ft buffer) — 6 (avoids all wetland with a 100 ft buffer) 12 (avoids all wetland with a 150 ft buffer) 2 (less than 10% disturbance) 8 (less than 5% disturbance) 16 (avoids 90% of wetland with a 200 ft buffer) 12 (100% avoidance) 20 (restoring previously disturbed areas) 16 (protection against future disturbance) Fig. 7. A coverage area is defined by its connecting points (magenta colored), as is its prefered style through the “select style/color” window. Table 3 Detailed level of achievement scenarios of credits CR1.1 and CR1.2. Climate and Resilience Emissions CR1.1 Reduce Net Embodied Carbon CR1.2 Reduce Greenhouse Gas Emissions Improved 5 (At Least 5% Reduction) 8 (At Least 10% Reduction) Enhanced 10 (At Least 15% Reduction) 13 (At Least 25% Reduction) Superior 15 (At Least 30% Reduction) 18 (At least 50% Reduction) Conserving 20 (At Least 50% Reduction) 22 (100% Reduction) Restorative — 26 (Carbon Negative) Table 4 Detailed Type 2 length reduction level of achievement scenarios of credits CR1.1 and CR1.2. Climate and Resilience Emissions CR1.1 Reduce Net Embodied Carbon CR1.2 Reduce Greenhouse Gas Emissions Improved Enhanced Superior Conserving 5 (more than 50% increment) 8 (more than 100% increment) 10 (less than 50% increment) 13 (less than 100% increment) 15 (less than 30% increment) 18 (less than 50% increment) 20 (less than 15% increment) 22 (less than 25% increment) another way of illustrating how these scenarios translate into achievable credit scores. CrCR1.1 5, ln ≥ 0.5 lu 10, 0.5 lu < ln ≤ 0.3 lu ={ 15, 0.3 lu < ln ≤ 0.15 lu 20, 0.15 lu < ln ≤ 0.05 lu 8, ln ≥ 2 lu 13, 2 lu < ln ≤ 0.5 lu CrCR1.2 = { 18, 0.5 lu < ln ≤ 0.25 lu 22, 0.25 lu < ln ≤ 0.1 lu (6) Restorative — 26 (Carbon Negative) (7) where lu is the Euclidean distance between the start- and end-points of the baseline alignment. Therefore, the objective function consists of the selected credits from Envision. It means that the generated candidates are evaluated based on 10 A. Laali et al. Sustainable Cities and Society 84 (2022) 104013 their fitness (compliance) regarding the credits’ constraints. Thus, the objective function is a total credit maximization (Crtotal ) function comprising the four credits, NW1.2 (CrNW1.2 ), NW1.3 (CrNW1.3 ), CR1.1 (CrCR1.1 ), and CR1.2 (CrCR1.2 ) presented in Eq. (9): Maximize Crtotal = CrNW1.2 + CrNW1.3 + CrCR1.1 + CrCR1.2 randomly picked to exchange genetic information between them (Ahmad Al-Hadad, 2011). This procedure forms two offspring, merging two separate segments from the parents represented in Fig. 9. 3.6.4.3. Mutation. Two mutation methods were used to help with evolving the performance of the solutions over successive generations. The mutation operator works on one individual simultaneously, maintaining diversity at a bit position while ensuring each individual’s mutability. A standard GA mutation operator is used for randomly changing the position of a randomly picked PI of an individual within the boundaries of the search space and the boundaries of the former and the subsequent PIs’ X coordinates. This will prevent the creation of unwanted loops in the produced alignments. If Xi is the coordinate value at the i position before the mutation, the operator assigns the new X value as (Eq. (12)): (8) 3.5.3. Initial population generation A random initial population of size is generated such that: • The start- and end-PIs define the boundary of the search area, which is determined by the BIM operator (design team) by using the input road smart object for creating the start- and end-points of the alignment (Fig. 8); • All PIs are within the search area’s boundaries so that each gene of a chromosome is assigned random X and Y coordinate values, as expressed in Eqs. (9) and (10): Xmin ≤ Xi ≤ Xmax (9) Ymin ≤ Yi ≤ Ymax (10) Xp = Xi−1 + rnd(Xi+1 − Xi−1 ) Furthermore, Grouped Point Mutation (GPM) dealt with a group of sequentially linked PIs (Al-Hadad, 2011), taking a bigger step in the mutation process while improving the alignments’ smoothness by replacing the sequential PIs with straight segments. The first generation of the alignments is generated based on their X values to produce candidates with fewer loops, such that (Eq. (11)): (11) Xi ≤ Xi+1 (12) 3.5.4. Selection and offspring generation 3.6.4.1. Selection. This study chooses GA’s parent chromosomes based on their fitness score for breeding the next generation. Crossover and mutation are then carried out on the selected parents for producing offspring. The offspring is then evaluated to find the fittest individuals and eliminate the more imperfect solutions for breeding the next generation by merging them with genetic operators. 3.6.4.2. Crossover. A single-point crossover method was used in this paper. This approach swaps genes from both parents in a single point Fig. 9. Crossover. Fig. 8. Determining the boundary of the search area using planning road smart objects in the BIM platform. 11 A. Laali et al. Sustainable Cities and Society 84 (2022) 104013 3.5.5. Stopping condition Two elimination criteria are used to stop the search for optimum solutions. The first one is reaching up to the total generation number. The end-user specifies this criterion, which is supposed to be big enough for the optimization algorithm to reach a meaningful and acceptable answer. The second criterion is if the best individual’s fitness remains unimproved for several successive generations. The user also defines the number of unimproved fitness scores before termination. The BIM-based sustainability algorithm enables designers and decision-makers to provide optimized sustainability alternatives with a detailed breakdown of the final presented solution’s performance concerning the defined credits from the sustainability assessment tool. Also, a 3D representation of the road alignment is generated. As mentioned, another advantage of the developed BIM-based optimization prototype is that the 3D model can automatically inherit certain properties of the utilized smart object for the optimization problem. This trait eliminates the need for the operator to redefine these properties, including the horizontal curve radius, minimum transition curve length, materials, and feature styles. This study selected planning roads to implement the developed prototype. Other input variables that can be adjusted based on the users’ preferences include population size, PI number, and total generation number. 4.1. Optimized alignments Fig. 10 represents the BIM model of the depicted study area with the defined boundary of the search area by the start- and end-points of the baseline alignment. The environmentally sensitive areas were imported in two steps. Two of these areas represent the same type of control area, i.e., conserved wetlands and one preserved prime farmland. The coverage area type properties are used for calculations in the script. The first obtained optimized alignment with a run time of 15 min was conducted with 8 PIs and a population size of 300 and 300 generations. The optimized alignment successfully avoids two environmentally sensitive areas. It inevitably crosses less than 515 m2s of one of the wetland conservation areas, gaining only 2 points from credit NW1.2. The total achieved score is 54, representing the best possible trade-offs among the four intended credits that provide an acceptable length reduction while keeping the required distance from control areas in a smaller window of opportunity than the former obtained optimized alignment. The second optimized alignment was conducted through 500 generations and a population size of 200 with 8 PIs, which took only about 24 min to obtain. This alignment’s fitness value is 65, which is achieved due to avoiding all of the environmentally valuable areas while limiting the road length to a minimum. Table 5 shows a detailed report of achieved points from each credit. The BIM operator can modify the output alignment and reevaluate the modified alignment using the sustainability assessment algorithm to produce another report on the alignment’s sustainability performance. The rest of the proposed alignments and their related models in Infraworks are presented in Table 5 and Fig. 11. 4. Model application This section tests the proposed framework’s applicability by testing the developed prototype on a hypothetical case study. The prototype was scripted using the latest Autodesk InfraWorks - JavaScript API Documentation 2014 to access the software’s internal data protocols. The Application Programming Interface (API) documentation is a technical source that guides how an API should be used and successfully integrated with the new application. This exhaustive manual (API documentation) covers all the required information to work with the API and provides information on functions, classes, forms of returns, arguments, and more. The prototype’s first development step was integrating the sustainability assessment criteria rules into the BIM platform to utilize the assessment algorithm as both the constraints and the optimization algorithm’s objective function. The sustainability assessment and optimization algorithm scripts are loaded into the scripting console. For the script to run, the user must select the baseline alignment and the environmentally sensitive areas imported as coverage areas into the model. The scripts use the selected objects (i.e., roads and environmentally sensitive areas) as inputs and then analyze the data to produce outputs in terms of a brief report on the achieved points with a 3D optimized and modifiable BIM road alignment model. However, sustainability involves various aspects and issues that entangle the design with critical economic, environmental, social, and geometric constraints. These constraints bind the road alignment completely, making the design process much more intricate. The achievable scores are dependent on the asset’s interaction with its surrounding environment and its improvements compared to the baseline design alternative. Accordingly, a model consisting of the surrounding environment and baseline model was created using the Model Builder feature in the BIM platform. The remaining data that cannot be sourced originally from the Model Builder feature can be imported into the model. In this study, the externally sourced data for evaluating credits NW1.2 and NW1.3 were the environmentally sensitive areas imported using shapefile formats. The defined rules and assessment criteria were then turned into scripted scenarios that calculate the level of achievement. Using JavaScript programming language, the scripted scenarios treat each object as a GeoJSON object and then calculate the locationdependent credits’ performance level. The achievement calculation level is based on the distance and overlaps of the road and the coverage areas. In contrast, in the case of length-dependent credits, the calculations are based on the suggested model’s percentage of improvement compared to the baseline model. 4.2. Sensitivity analysis To illustrate the model capabilities and further validation of the developed prototype, an extensive analysis of sensitivity to key model parameters is performed. The performed sensitivity analyses of the proposed model are based on the study conducted by Kang et al. (2012) and done by excluding or changing the conditioning scenarios and factors. 4.2.1. Sensitivity to objective function components The sensitivity analysis to objective function components is intended to demonstrate the model formulation’s sensitivity to the credits selected from Envision. It shows these credits’ applicability to road Fig. 10. Case study area and baseline inputs. 12 A. Laali et al. Sustainable Cities and Society 84 (2022) 104013 Maximize Crtotal = CrCR1.1 + CrCR1.2 + 300 500 500 300 500 generations Population size 200 200 200 200 200 alignment design optimization and their relevance for constructing a highway optimization model. Three different scenarios, as presented in Eqs. (13), (14), and (15) were designed to show each evaluation criteria’s impact on the resulting road alignments. All input parameters used in the scenarios are identical except for the credits selected to compose the objective function as follows: Scenario A1 Scenario A2 runtime (Seconds) 366.479 1440.74 4444.458 3000 1386.213 Maximize Crtotal = CrNW1.2 + CrNW1.3 Maximize Crtotal = CrNW1.2 + CrNW1.3 + CrCR1.1 + CrCR1.2 Length increment percentage 13.81% 22.62% 17.92% 31.80% 18.89 Disturbance (m2 ) 515.3 0 0 0 0 Alignment length 15,202 16,377.8 15,750 17,604 15,880 4.2.2. Sensitivity to length-dependent scenario design Due to this study’s effort to integrate Envision with Infraworks to facilitate sustainability assessment and optimization, the scenarios were defined precisely based on the evaluation systems instructions for evaluation. However, each criterion in Envision consists of only 4 or 5 scenarios. Therefore, the alignment length reduction scenarios have large intervals that undermine the optimization algorithm’s effort to reduce the path length. If these scenarios are broken down to a greater extent and placed in more precise and rigorous intervals, such as Table 7, the path length reduction will improve. It should be noted that breaking the final scenarios into smaller intervals had an improved effect on reducing the length. Since Envision was considered the base sustainability assessment tool for implementing the proposed framework, the evaluation scenarios were formulated strictly based on the system’s defined levels of achievement evaluation criteria. Because these scenarios’ measurement sensitivity is regulated over a wide range, they impose a relatively smaller conditioning effect on the optimization problem. For instance, the difference in reduction percentage between the superior and conserving levels of achievement in credit CR1.1 is 20%, which leaves a wide range for the optimization algorithm in length reduction; thus, the length reduction was between 30% and 50%, which is a very wide range and can acquire a high score. Therefore, if the scenarios were broken down into smaller, more strict intervals, as presented in Table 7. The optimization algorithm’s effort for length reduction improved based on the Type 2 length reduction scenario design. Comparing the result alignments in Fig. 13, Case 3-A (optimized with scenarios originally defined by Envision) and Fig. 13, Case 3-A (optimized through more restrictive length reduction scenarios) makes this matter evident. The results are presented in Table 8. 12 12 12 12 12 2 16 16 16 16 20 15 15 10 15 22 22 22 18 22 Total score 54 65 65 56 65 CR1.2 CR1.1 NW1.3 (15) The alignment optimized with scenario A1, while disregarding the location-dependent credits in the formulation of the objective function, as shown in Fig. 12 Case 2-B, has a length of about 8.82 miles that affects high environmental value areas. Fig. 12 Case 2-B shows the optimized alignment with locationdependent credits as the objective function. As shown in Table 6, Case 2-B, with a length of 11 miles, has not affected any environmentally sensitive areas or buffers. However, the route’s length is about 1.2 miles longer than in the third case (Fig. 12 Case 2-C), which includes both the length reduction and high environmental valued site avoidance credits as the optimization problem’s objective, as shown in Table 6. PI number 8 8 15 15 8 Length reduction scenario type 1 1 2 1 2 NW1.2 (14) Scenario A3 4.2.3. Sensitivity to PI number Determining the optimal number of PIs is a key input for the model implementation in individual projects. As such, this parameter may affect the accuracy of the solutions. In this section, different scenarios are designed to demonstrate the proposed model’s sensitivity to the number of PIs. Also, the model’s runtime may vary by changing this parameter. To discover the optimal number of PIs in road alignment Optimized alignment Case 1 Case 2 Case 3 Case 4 Case 5 Table 5 Detailed report of the BIM-based optimized alignments achieved points from each credit. (13) 13 A. Laali et al. Sustainable Cities and Society 84 (2022) 104013 Fig. 11. Optimized alignments. Fig. 12. Sensitivity of optimized alignments to objective function components. consists of only two categories of credit components, no more than 15 PIs were tested. As the results indicate, increasing the number of PIs can improve the final obtained alignments with increased population size and generation size because of the increased search opportunity and the added precision. It is also observed that a greater number can overburden the computation process and increase the runtime while causing dysfunction. Table 6 Sensitivity analysis results based on objective function components. Scenario Length increment Alignment length Score from CR1.2 Score from CR1.1 Total score A1 A2 A3 6.63% 29.76% 18.89% 14,242 17,331.3 15,880 22 18 22 20 15 15 42 33 37 5. Discussion optimization, we ran the model several times with different numbers of PIs. Fig. 14 shows the optimized alignments with varying sizes of population and PI numbers through different generation numbers (5, 10, and 15). The detailed results of each case are presented in Table 9. Since the search area is relatively small and the objective function This BIM-based end-user application intends to simplify and speed up the process of finding desirable sustainable infrastructure alternatives. At the same time, it will encourage the implementation of sustainable measures and assist designers in making project-related 14 A. Laali et al. – 20 (up to 5% length increase) 20 (up to 15% length increase) 17 (up to 10% length increase) 18 (up to 20% length increase) 22 (up to 5% length increase) Fig. 13. Sensitivity to more restricted scenario design. 14 (up to 15% length increase) 16 (up to 25% length increase) Scenario 7 2 (more than 50% length increase) 8 (more than 100% increase) CR1.1 Reduce Net Embodied Carbon CR1.2 Reduce Greenhouse Gas Emissions 5 (up to 50% length increase) 10 (up to 100% length increase) 8 (up to 35% length increase) 12 (up to 65% length increase) 11 (up to 25% length increase) 14 (up to 40% length increase) decisions. The results from the model application to a hypothetical project in five cases (Table 9) indicate that first, with the increase of PI numbers and generations, the runtime also increases. Simultaneously, the optimized alignments also significantly satisfy higher performance achievement levels. The sensitivity objective function components analysis confirms that sustainability evaluation criteria can and should be jointly optimized for effective optimization. Many trade-off opportunities exist depending on the flexibility and preferences specified with the input parameters. Furthermore, although it may seem obvious, a rise in PI numbers can benefit from an elevated generation size to yield more accurate and refined results. The developed prototype allows for repetitive real-time sustainability assessment of infrastructure projects at various design stages without the use of additional software. Furthermore, it has been shown that the model can not only effectively optimize highway alignments that satisfy the chosen sustainability evaluation criteria and provide information about the resulting alignments to designers but also satisfy basic highway design standards. This is a result of the BIM-based optimized model with its parametric capabilities (e.g., predefined in compliance with road standards and real-time adaptation of territorial conditions). Additionally, the optimum alignments were obtained without difficulty and within reasonable calculation timeframes. It’s worth noting that these results resonate with what used to be merely a theory. The model is expected to apply to many real-world projects (other types of projects) and perform well in finding the most desirable alternative alignments during the early stages of road development. While the framework and prototype are generalizable and customizable, the breadth of the pilot implementation was limited, and several considerations could be extensively explored in future studies. Therefore, despite the established strengths, the subsequent model enhancements are desirable: • Although this model does not cover all of Envision’s credits in the prototype, it provides a foundation for further development and an expansion of BIM processes towards sustainability issues. The optimization problem’s objective function comprises only four sustainability criteria from the baseline sustainability evaluation system: NW1.2, NW1.3, CR1.1, and CR1.2. However, all Envision credits must be considered to build comprehensively balanced sustainable alternatives. For instance, adding credit “RA1.5 Balance Earthwork on Site” to the evaluation criteria can significantly improve the final outcome as it responds to both sustainability considerations and general design objectives. This is because this credit intends to minimize transportation and environmental consequences by minimizing the transportation of soils and other excavated items off-site. CLIMATE AND RESILIENCE EMISSIONS Scenario 1 Credits Table 7 Design with more restrictive evaluation scenarios for length reduction credits. Scenario 2 Scenario 3 Scenario 4 Scenario 5 Scenario 6 Scenario 8 Sustainable Cities and Society 84 (2022) 104013 15 A. Laali et al. Sustainable Cities and Society 84 (2022) 104013 Table 8 Sensitivity analysis results based on scenario design of length reduction credits. Optimized alignment Alignment length Length reduction type Length increment percentage Runtime (seconds) Generations Population size Case 3-A Case 3-B 14,244 13,592 1 2 6.64% 1.76% 24.247 36.468 500 500 200 200 Fig. 14. Sensitivity to PI number. At the same time, the BIM tool can reciprocate through the production of realistic and computable terrain representations. • In parallel to this, to ensure not only sustainability but also costeffectiveness, it is desirable to improve the model’s evaluation criteria for yielding more accurate and realistic outcomes. Thus, in an upgraded model version, cost-effectiveness should be simultaneously evaluated and optimized for all the generated highway alignments through a trade-off analysis. • The uniqueness of diverse infrastructure projects necessitates specialization of the framework’s elements down to the last details of each of them. This means that specific decision variables of each infrastructure type determine how each framework element is employed. For instance, although Envision is available to all forms of 16 Sustainable Cities and Society 84 (2022) 104013 A. Laali et al. Table 9 PI sensitivity analysis results. Optimized alignment PI number Alignment length Total score Runtime (seconds) Generations Population size Case 1-A Case 1-B Case 1-C Case 1-D Case 1-E Case 1-F Case 1-G 5 5 5 10 10 15 15 17,593 15,302.1 15,145 14,687 1609.5 16,385 15,697.7 56 56 56 46 70 65 65 2743.360 2555.214 1596.850 1807.147 2743.360 3154.568 4345.724 200 300 200 200 300 200 300 300 500 500 300 500 300 500 critical hindrances to BIM-based infrastructure sustainability improvement—initial and additional costs due to changes in traditional work processes and sustainability analyses. In conclusion, this paper demonstrated the plausibility of BIM and Envision integration for automated sustainability assessment and optimization under certain conditions. The developed prototype uses a BIM platform with smart objects designed for roads. However, BIM platforms might not be capable of modeling other types of infrastructure as smart objects. In that case, the capabilities of BIM should be extended in future works to include different types of impactful infrastructure and possibly, enable a systemic and interrelated analysis of interconnected infrastructure systems. infrastructure, the performance evaluation criteria of its credits may differ for each infrastructure type, or they may not apply at all. While each element or the detailed process of their use might be altered, the overall data flow can still follow the proposed framework. For example, if PSO were selected instead of GA, the optimization procedure would have changed, but the overall information flow would still follow Fig. 1. • The results are limited to Envision as the baseline for performing the optimization and assessment algorithms. Envision determines the interlinkage between the sustainability credits and their relative weight in the current version of the model, which is a key input parameter affecting the generated highway alignments. It affects the trade-off between the corresponding objective function components and, consequently, the outcome. Hence, the balance between the applicable criteria needs to be in line with the context of the potential project. Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Hence, further extensions of this research might include modeling other sustainability criteria, cost components, or decision objectives to increase the optimization’s reliability and improve its final presentable solutions. Acknowledgments This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. 6. Conclusions This paper proposes a BIM-based framework for automating sustainability assessment and optimizing infrastructure projects. This framework was tested through a prototype development used in a case study to prove its reliability, effectiveness, and efficiency. A sensitivity analysis of the key model parameters was then performed to test the model’s reliability and Envision’s selected credits concerning an alignment optimization problem. The proposed automated framework makes it possible for designers to investigate and understand the potential sustainability impact of an infrastructure project in real time during the design stage. Notably, these projects have wide-ranged sustainability impacts at a time of expected expansion by almost two and a half times due to population growth and the increasing importance of sustainability issues. Additionally, the developed model uses a BIM platform for automating both the assessment and the optimization algorithm; therefore, project teams and sustainability assessment agencies can access the BIM model in a single platform. Hence, team projects can provide the required sustainability evaluation documentation within a shareable environment. Also, sustainability assessment agencies can use the BIMbased assessment algorithm to qualify the BIM-based documentation provided by the project team and certify infrastructure projects in an efficient and timely manner. Moreover, the developed prototype avoids interoperability issues by creating entirely BIM-based algorithms. The integrated BIM process classifies the BIM object’s attributes and facilitates the extraction of values demanded by the rating system and thus the sustainability evaluation of any given infrastructure alternative. Another key contribution is the reduced cognitive load of balancing the complicated tradeoff between conflicting infrastructure sustainability performance criteria through the automated optimization of positive sustainability impact. Also, the proposed framework can overcome two of the most Supplementary materials Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.scs.2022.104013. References Adshead, D., Thacker, S., Fuldauer, L. I., & Hall, J. W. (2019). Delivering on the sustainable development goals through long-term infrastructure planning. Global Environmental Change, 59, Article 101975. AGC, A. (2005). The contractors guide to BIM. URL: http://iweb.agc.org/iweb/Purchase /ProductDetail.aspx. Ahmad Al-Hadad, B. (2011). An approach to the highway alignment development process using genetic algorithm based optimisation. University of Nottingham. Ainger, C. M., & Fenner, R. A. (2014). Sustainable infrastructure: Principles into practice. London, UK: ICE publishing. Al-Hadad, B. A. (2011). An approach to the highway alignment development process using genetic algorithm based optimisation. University of Nottingham. Al-Hadad, B. M. A., & Mawdesley, M. (2010). A genetic algorithm approach to a 3D highway alignment development. IJCCI (ICEC), 129–136. Ando, S., Arima, T., Bogaki, K., Hasegawa, H., Hoyano, A., & Ikaga, T. (2005). Architecture for a sustainable future. Tokyo: Architectural Institute of Japan. Azhar, S., Carlton, W. A., Olsen, D., & Ahmad, I. (2011). Building information modeling for sustainable design and LEED® rating analysis. Automation in Construction, 20, 217–224. Basbagill, J., Flager, F., Lepech, M., & Fischer, M. (2013). Application of life-cycle assessment to early stage building design for reduced embodied environmental impacts. Building and Environment, 60, 81–92. Biswas, T., Wang, T., & Sung-Hsien, R., K.. (2008). Integrating sustainable building rating systems with building information models. Bongiorno, N., Bosurgi, G., Carbone, F., Pellegrino, O., & Sollazzo, G. (2019). Potentialities of a highway alignment optimization method in an I-BIM environment. Periodica Polytechnica Civil Engineering, 63, 352–361. Bradley, A., Li, H., Lark, R., & Dunn, S. (2016). BIM for infrastructure: An overall review and constructor perspective. Automation in Construction, 71, 139–152. Bueno, C., & Fabricio, M. M. (2018). Comparative analysis between a complete LCA study and results from a BIM-LCA plug-in. Automation in Construction, 90, 188–200. 17 Sustainable Cities and Society 84 (2022) 104013 A. Laali et al. Carvalho, J. P., Bragança, L., & Mateus, R. (2019). Optimising building sustainability assessment using BIM. Automation in Construction, 102, 170–182. Chan, D. W., Olawumi, T. O., & Ho, A. M. (2019). Perceived benefits of and barriers to Building Information Modelling (BIM) implementation in construction: The case of Hong Kong. Journal of Building Engineering, 25, Article 100764. Chong, H. Y., Lopez, R., Wang, J., Wang, X., & Zhao, Z. (2016). Comparative analysis on the adoption and use of BIM in road infrastructure projects. Journal of Management in Engineering, 32, Article 05016021. Clevenger, C. M., Ozbek, M. E., & Simpson, S. (2013). Review of sustainability rating systems used for infrastructure projects. In 49th ASC annual international conference proceedings (pp. 10–13). Asscociated Schools for Construction. Cole, J. (2003). Building environmental assessment methods-a measure of success. Construction. Citeseer. Costin, A., Adibfar, A., Hu, H., & Chen, S. S. (2018). Building Information Modeling (BIM) for transportation infrastructure–Literature review, applications, challenges, and recommendations. Automation in Construction, 94, 257–281. Diaz-Sarachaga, J. M., Jato-Espino, D., Alsulami, B., & Castro-Fresno, D. (2016). Evaluation of existing sustainable infrastructure rating systems for their application in developing countries. Ecological Indicators, 71, 491–502. Chen, S.Y., Lok, K., & Jeng, T. (2016). Smart BIM objects for design intelligence. Envision, (2020). Envision [Online]. Available: https://sustainableinfrastructure.org/ [Accessed]. Ferrer, A. L. C., Thomé, A. M. T., & Scavarda, A. J. (2018). Sustainable urban infrastructure: A review. Resources, Conservation and Recycling, 128, 360–372. French, G. (2012). State of the world report 2012: Sustainable infrastructure. International Federation of Consulting Engineers (FIDIC). Gaughan, M. (2012). Assessing sustainability of pipeline projects using envision rating system. Pipelines 2012: Innovations in design, construction, operations, and maintenance, doing more with less. Geyer, P. (2009). Component-oriented decomposition for multidisciplinary design optimization in building design. Advanced Engineering Informatics, 23, 12–31. BreGlobal. (2021). Ceequal. In B. Global (Ed.), BRE global launches ceequal version 6 to support infrastructure projectsteams sustainability goals. ed. Manual available for purchase at website entitled. Griffiths, K. A., Boyle, C., & Henning, T. F. (2017). In Comparative assessment of infrastructure sustainability rating tools. Conference: Transportation Research Board 96th Annual Meeting-Transportation Research Board (p. 01265). Hartmann, T., Van Meerveld, H., Vossebeld, N., & Adriaanse, A. (2012). Aligning building information model tools and construction management methods. Automation in Construction, 22, 605–613. Hergunsel, M.F. (.2011). Benefits of building information modeling for construction managers and BIM based scheduling. Hollberg, A., Genova, G., & Habert, G. (2020). Evaluation of BIM-based LCA results for building design. Automation in Construction, 109, Article 102972. Hooper, M., & Ekholm, A. (2012). A BIM-Info delivery protocol. Construction Economics and Building, 12, 39–52. Hurley, J. (2009). Sustainable or status-quo: Investigating sustainability assessment of residential estate development. In 4th State of Australian Cities National Conference, 24-27 November 2009 …. IDB. (2018). What Is Sustainable Infrastructure?. A framework to guide sustainability across the project cycle. Washington, DC: Inter-American Development Bank. IDB Invest. Infrastructure, I. F. S., & Infrastructure, Z. P. F. S. (2018). Envision sustainable infrastructure framework, version 3. Washington, DC: Institute for Sustainable Infrastructure. ISI. (2018b). In I. F. S. Infrastructure (Ed.), Envision: Sustainable infrastructure framework guidance manual (3rd ed.). Washington, DC: Institute for Sustainable Infrastructure. ISCA. (2021). IS Rating Scheme [Online]. Available: Available at https://www.cer tifiedenergy.com.au/is-rating [Accessed 01/06 2021]. ISI. (2015). Envision rating system for sustainable infrastructure. Washington, DC, USA: Institute for Sustainable Infrastructure. ISI. (2018a). Envision. Washington, DC: Institute for Sustainable Infrastructure. ISI. (2021). Envision [Online]. Available: Https://sustainableinfrastructure.org/resource /website-tutorials/ [Accessed 06 /01 2021]. Jha, M. K., & Schonfeld, P. (2000). Geographic information system–based analysis of right-of-way cost for highway optimization. Transportation Research Record, 1719, 241–249. Ji, Y., Borrmann, A., Beetz, J., & Obergrießer, M. (2013). Exchange of parametric bridge models using a neutral data format. Journal of Computing in Civil Engineering, 27, 593–606. Kang, M.-W., Jha, M. K., & Schonfeld, P. (2012). Applicability of highway alignment optimization models. Transportation Research Part C: Emerging Technologies, 21, 257–286. Kim, J.-U., Kim, Y.-J., Ok, H., & Yang, S.-H. (2015). A study on the status of infrastructure BIM and BIM library development. In 2015 International Conference on Computational Science and Computational Intelligence (CSCI) (pp. 857–858). IEEE. Liu, Y., Van Nederveen, S., Wu, C., & Hertogh, M. (2018). Sustainable infrastructure design framework through integration of rating systems and building information modeling. Advances in Civil Engineering, 2018. Liu, Z., LU, Y., Shen, M., & Peh, L. C. (2020). Transition from building information modeling (BIM) to integrated digital delivery (IDD) in sustainable building management: A knowledge discovery approach based review. Journal of Cleaner Production, Article 125223. Lorek, S. (2018). What is bim (building information modeling). Maji, A., & Jha, M. K. (2009). Multi-objective highway alignment optimization using a genetic algorithm. Journal of Advanced Transportation, 43, 481–504. Marzouk, M., & Othman, A. (2020). Planning utility infrastructure requirements for smart cities using the integration between BIM and GIS. Sustainable Cities and Society, 57, Article 102120. Mcvoy, G.R., Nelson, D.A., Krekeler, P., Kolb, E., & Gritsavage, J.S. (.2010).Moving towards Sustainability: New York State Department of Transportation’s GreenLITES Story. Green streets and highways 2010: An interactive conference on the state of the art and how to achieve sustainable outcomes,461–479. Minagawa, M., & Kusayanagi, S. (2015). Study on BIM utilization for design improvement of infrastructure project. Procedia engineering, 125, 431–437. Nguyen, A.-T., Reiter, S., & Rigo, P. (2014). A review on simulation-based optimization methods applied to building performance analysis. Applied Energy, 113, 1043–1058. Rahmani Asl, M., Zarrinmehr, S., Bergin, M., & Yan, W. (2015). BPOpt: A framework for BIM-based performance optimization. Energy and Buildings, 108, 401–412. Rees, W. (2010). What’s blocking sustainability? Human nature, cognition, and denial. Sustainability: Science, Practice and Policy, 6, 13–25. Sabatino, S., Frangopol, D. M., & Dong, Y. (2015). Sustainability-informed maintenance optimization of highway bridges considering multi-attribute utility and risk attitude. Engineering Structures, 102, 310–321. Santos, R., Costa, A. A., Silvestre, J. D., & PYL, L. (2020). Development of a BIM-based environmental and economic life cycle assessment tool. Journal of Cleaner Production, 265, Article 121705. Saville, C. R., Miller, G. R., & Brumbelow, K. (2016). Using envision to assess the sustainability of groundwater infrastructure: A case study of the twin oaks aquifer storage and recovery project. Sustainability, 8, 501. Schmid, S., Galicz, E., & Reinhardt, W. (2015). WMS performance of selected SQL and NoSQL databases. In International Conference on Military Technologies (ICMT) 2015 (pp. 1–6). IEEE. Shahtaheri, Y., Flint, M. M., & De LA Garza, J. M. (2018). Sustainable infrastructure multi-criteria preference assessment of alternatives for early design. Automation in Construction, 96, 16–28. Thangaraj, R., Pant, M., Abraham, A., & Bouvry, P. (2011). Particle swarm optimization: Hybridization perspectives and experimental illustrations. Applied Mathematics and Computation, 217, 5208–5226. Tryggestad, K. (2013). Sustainable buildings and infrastructure: Paths to the future. Taylor & Francis. Tsai, C. Y., & Chang, A. S. (2012). Framework for developing construction sustainability items: The example of highway design. Journal of Cleaner Production, 20, 127–136. UN. (2015). Transforming our world: The 2030 agenda for sustainable development. New York: United Nations, Department of Economic and Social Affairs. USDOT. (2018). Highway Statistics [Online]. Available: Https://www.fhwa.dot.gov/po licyinformation/statistics/2017/hm260.cfm [Accessed 2021]. van Eldik, M. A., Vahdatikhaki, F., Dos Santos, J. M. O., Visser, M., & Doree, A. (2020). BIM-based environmental impact assessment for infrastructure design projects. Automation in Construction, 120, Article 103379. WCED, (1987). Report of the World Commission on Environment and Development: Our common future. Accessed Feb, 10. Willetts, R., Burdon, J., Glass, J., & Frost, M. (2010). Environmental and sustainability impact assessment of infrastructure in the United Kingdom. Transportation research record, 2158, 143–150. Wong, J. K. W., & Zhou, J. (2015). Enhancing environmental sustainability over building life cycles through green BIM: A review. Automation in Construction, 57, 156–165. 18