This paper introduces an extension of the multi-start local search framework [1, 2, 3], named Sys... more This paper introduces an extension of the multi-start local search framework [1, 2, 3], named Systematic Diversification for the Vehicle Routing Problem with Time Windows, where the objective is to design least cost routes for a fleet of vehicles from one depot to a set of geographically scattered points. The routes must be designed in such a way that each point is visited only once by exactly one vehicle within a given time interval; all routes start and end at the depot, and the total demands of all points on one particular route must not exceed the capacity of the vehicle. The problem has been under intensive research during the past few years because of its high practical importance. The suggested method is based on systematically penalizing solution features/attributes to diversify the search within a multi-start local search, where the basic idea is to create several initial solutions, and apply improvement heuristics to each of them. Then, the best solution found during the search is the final output. The experimental results on the well-known 100-customer benchmarks show that the described approach is reliable, effective and robust, being very competitive with all previous methods found in the literature.
In this paper we describe synergy effects of combining state-of-the-art Geographic Information Te... more In this paper we describe synergy effects of combining state-of-the-art Geographic Information Technology (GIT) with novel methods for planning and scheduling from the field of Constraint Reasoning (CR). We present a method for solving what we have called the Clear-Cut Scheduling Problem (CCSP), where the task is to assign clear-cutting times to regions in a given forest area over a
This report surveys the research on the metaheuristics for the Vehicle Routing Problem with Time ... more This report surveys the research on the metaheuristics for the Vehicle Routing Problem with Time Windows (VRPTW). The VRPTW can be described as the problem of designing least cost routes from one depot to a set of geographically scattered points. The routes must be designed in such a way that each point is visited only once by exactly one vehicle
A typical vehicle routing problem can be described as the problem of designing least cost routes ... more A typical vehicle routing problem can be described as the problem of designing least cost routes from one depot to a set of geographically scattered points (cities, stores, warehouses, schools, customers etc). The routes must be designed in such a way that each point is visited only once by exactly one vehicle, all routes start and end at the depot,
... For a focused literature survey, we refer to Flatberg et al. ... the solver must support lock... more ... For a focused literature survey, we refer to Flatberg et al. ... the solver must support locking parts of the current plan that should not be modified during further optimization. ... The inherent complexity of these problems has led us to choose a unified meta-heuristic approach in SPIDER ...
In many cases there is still a large gap between the performance of current optimization technolo... more In many cases there is still a large gap between the performance of current optimization technology and the requirements of real-world applications. As in the past, performance will improve through a combination of more powerful solution methods and a general performance increase of computers. These factors are not independent. Due to physical limits, hardware development no longer results in higher speed for sequential algorithms, but rather in increased parallelism. Modern commodity PCs include a multi-core CPU and at least one GPU, providing a low-cost, easily accessible heterogeneous environment for high-performance computing. New solution methods that combine task parallelization and stream processing are needed to fully exploit modern computer architectures and profit from future hardware developments. This paper is the second in a series of two. Part I gives a tutorial style introduction to modern PC architectures and GPU programming. Part II gives a broad survey of the literature on parallel computing in discrete optimization targeted at modern PCs, with special focus on routing problems. We assume that the reader is familiar with GPU programming, and refer the interested reader to Part I. We conclude with lessons learnt, directions for future research, and prospects.
In many cases there is still a large gap between the performance of current optimization technolo... more In many cases there is still a large gap between the performance of current optimization technology and the requirements of real-world applications. As in the past, performance will improve through a combination of more powerful solution methods and a general performance increase of computers. These factors are not independent. Due to physical limits, hardware development no longer results in higher speed for sequential algorithms, but rather in increased parallelism. Modern commodity PCs include a multi-core CPU and at least one GPU, providing a low-cost, easily accessible heterogeneous environment for high-performance computing. New solution methods that combine task parallelization and stream processing are needed to fully exploit modern computer architectures and profit from future hardware developments. This paper is the second in a series of two. Part I gives a tutorial style introduction to modern PC architectures and GPU programming. Part II gives a broad survey of the literature on parallel computing in discrete optimization targeted at modern PCs, with special focus on routing problems. We assume that the reader is familiar with GPU programming, and refer the interested reader to Part I. We conclude with lessons learnt, directions for future research, and prospects.
The purpose of this paper is to present a new deterministic annealing metaheuristic for the fleet... more The purpose of this paper is to present a new deterministic annealing metaheuristic for the fleet size and mix vehicle routing problem with time windows. The objective is to service, at minimal total cost, a set of customers within their time windows by a heterogeneous capacitated vehicle fleet. First, we motivate and define the problem. We then proceed to
... Department of Industrial Economics and Technology Management, Norwegian University of Science... more ... Department of Industrial Economics and Technology Management, Norwegian University of Science and Technology, NO-7491 Trondheim, Norway, marielle ... Since 1991, the TRISTAN conferences have become one of the main gathering points of top researchers from ...
... AUTHOR(S) Truls Flatberg, Geir Hasle, Oddvar Kloster, Eivind J. Nilssen, and Atle Riise CLIEN... more ... AUTHOR(S) Truls Flatberg, Geir Hasle, Oddvar Kloster, Eivind J. Nilssen, and Atle Riise CLIENT(S) ... Examples are time dependent travel times in the road network (eg rush hour congestions), and time dependent probabilities for occurrence of new transport requests in a ...
There is a strong pressure on economy in the distribution of media products. Two important remedi... more There is a strong pressure on economy in the distribution of media products. Two important remedies are more efficient carrier routes and distribution of side products. Both call for effective and dynamic route design and revision processes. These processes are complex, time-consuming, and costly. The size of industrial carrier route planning instances may cause performance problems for VRP algorithms. In the VRP literature, the Capacitated Arc Routing Problem (CARP) is often advocated as an adequate model for applications such as newspaper delivery and garbage collection. We argue that a better model is the Node Edge Arc Routing Problem (NEARP). We describe how we have extended a VRP solver to enable modeling of the NEARP, and extended it with a framework for multi-level aggregation of demand. An aggregation heuristic that is based on the underlying road topology is presented. The resulting solver has been integrated in a commercial web based system for route management and tested by pilot users. We present experimental results on real-life data from newspaper distribution. Results on standard CARP and NEARP instances from the literature are given, including several new best known solutions.
... One may run multiple kernels concurrently. Page 22. Applied Mathematics GPU vs CPU performanc... more ... One may run multiple kernels concurrently. Page 22. Applied Mathematics GPU vs CPU performance Page 23. Applied Mathematics Programming GPUs ... Single die heterogeneous processors ∎ AMD Fusion ∎ Intel Sandy Bridge Page 30. Applied Mathematics Why bother? ...
This paper introduces an extension of the multi-start local search framework [1, 2, 3], named Sys... more This paper introduces an extension of the multi-start local search framework [1, 2, 3], named Systematic Diversification for the Vehicle Routing Problem with Time Windows, where the objective is to design least cost routes for a fleet of vehicles from one depot to a set of geographically scattered points. The routes must be designed in such a way that each point is visited only once by exactly one vehicle within a given time interval; all routes start and end at the depot, and the total demands of all points on one particular route must not exceed the capacity of the vehicle. The problem has been under intensive research during the past few years because of its high practical importance. The suggested method is based on systematically penalizing solution features/attributes to diversify the search within a multi-start local search, where the basic idea is to create several initial solutions, and apply improvement heuristics to each of them. Then, the best solution found during the search is the final output. The experimental results on the well-known 100-customer benchmarks show that the described approach is reliable, effective and robust, being very competitive with all previous methods found in the literature.
In this paper we describe synergy effects of combining state-of-the-art Geographic Information Te... more In this paper we describe synergy effects of combining state-of-the-art Geographic Information Technology (GIT) with novel methods for planning and scheduling from the field of Constraint Reasoning (CR). We present a method for solving what we have called the Clear-Cut Scheduling Problem (CCSP), where the task is to assign clear-cutting times to regions in a given forest area over a
This report surveys the research on the metaheuristics for the Vehicle Routing Problem with Time ... more This report surveys the research on the metaheuristics for the Vehicle Routing Problem with Time Windows (VRPTW). The VRPTW can be described as the problem of designing least cost routes from one depot to a set of geographically scattered points. The routes must be designed in such a way that each point is visited only once by exactly one vehicle
A typical vehicle routing problem can be described as the problem of designing least cost routes ... more A typical vehicle routing problem can be described as the problem of designing least cost routes from one depot to a set of geographically scattered points (cities, stores, warehouses, schools, customers etc). The routes must be designed in such a way that each point is visited only once by exactly one vehicle, all routes start and end at the depot,
... For a focused literature survey, we refer to Flatberg et al. ... the solver must support lock... more ... For a focused literature survey, we refer to Flatberg et al. ... the solver must support locking parts of the current plan that should not be modified during further optimization. ... The inherent complexity of these problems has led us to choose a unified meta-heuristic approach in SPIDER ...
In many cases there is still a large gap between the performance of current optimization technolo... more In many cases there is still a large gap between the performance of current optimization technology and the requirements of real-world applications. As in the past, performance will improve through a combination of more powerful solution methods and a general performance increase of computers. These factors are not independent. Due to physical limits, hardware development no longer results in higher speed for sequential algorithms, but rather in increased parallelism. Modern commodity PCs include a multi-core CPU and at least one GPU, providing a low-cost, easily accessible heterogeneous environment for high-performance computing. New solution methods that combine task parallelization and stream processing are needed to fully exploit modern computer architectures and profit from future hardware developments. This paper is the second in a series of two. Part I gives a tutorial style introduction to modern PC architectures and GPU programming. Part II gives a broad survey of the literature on parallel computing in discrete optimization targeted at modern PCs, with special focus on routing problems. We assume that the reader is familiar with GPU programming, and refer the interested reader to Part I. We conclude with lessons learnt, directions for future research, and prospects.
In many cases there is still a large gap between the performance of current optimization technolo... more In many cases there is still a large gap between the performance of current optimization technology and the requirements of real-world applications. As in the past, performance will improve through a combination of more powerful solution methods and a general performance increase of computers. These factors are not independent. Due to physical limits, hardware development no longer results in higher speed for sequential algorithms, but rather in increased parallelism. Modern commodity PCs include a multi-core CPU and at least one GPU, providing a low-cost, easily accessible heterogeneous environment for high-performance computing. New solution methods that combine task parallelization and stream processing are needed to fully exploit modern computer architectures and profit from future hardware developments. This paper is the second in a series of two. Part I gives a tutorial style introduction to modern PC architectures and GPU programming. Part II gives a broad survey of the literature on parallel computing in discrete optimization targeted at modern PCs, with special focus on routing problems. We assume that the reader is familiar with GPU programming, and refer the interested reader to Part I. We conclude with lessons learnt, directions for future research, and prospects.
The purpose of this paper is to present a new deterministic annealing metaheuristic for the fleet... more The purpose of this paper is to present a new deterministic annealing metaheuristic for the fleet size and mix vehicle routing problem with time windows. The objective is to service, at minimal total cost, a set of customers within their time windows by a heterogeneous capacitated vehicle fleet. First, we motivate and define the problem. We then proceed to
... Department of Industrial Economics and Technology Management, Norwegian University of Science... more ... Department of Industrial Economics and Technology Management, Norwegian University of Science and Technology, NO-7491 Trondheim, Norway, marielle ... Since 1991, the TRISTAN conferences have become one of the main gathering points of top researchers from ...
... AUTHOR(S) Truls Flatberg, Geir Hasle, Oddvar Kloster, Eivind J. Nilssen, and Atle Riise CLIEN... more ... AUTHOR(S) Truls Flatberg, Geir Hasle, Oddvar Kloster, Eivind J. Nilssen, and Atle Riise CLIENT(S) ... Examples are time dependent travel times in the road network (eg rush hour congestions), and time dependent probabilities for occurrence of new transport requests in a ...
There is a strong pressure on economy in the distribution of media products. Two important remedi... more There is a strong pressure on economy in the distribution of media products. Two important remedies are more efficient carrier routes and distribution of side products. Both call for effective and dynamic route design and revision processes. These processes are complex, time-consuming, and costly. The size of industrial carrier route planning instances may cause performance problems for VRP algorithms. In the VRP literature, the Capacitated Arc Routing Problem (CARP) is often advocated as an adequate model for applications such as newspaper delivery and garbage collection. We argue that a better model is the Node Edge Arc Routing Problem (NEARP). We describe how we have extended a VRP solver to enable modeling of the NEARP, and extended it with a framework for multi-level aggregation of demand. An aggregation heuristic that is based on the underlying road topology is presented. The resulting solver has been integrated in a commercial web based system for route management and tested by pilot users. We present experimental results on real-life data from newspaper distribution. Results on standard CARP and NEARP instances from the literature are given, including several new best known solutions.
... One may run multiple kernels concurrently. Page 22. Applied Mathematics GPU vs CPU performanc... more ... One may run multiple kernels concurrently. Page 22. Applied Mathematics GPU vs CPU performance Page 23. Applied Mathematics Programming GPUs ... Single die heterogeneous processors ∎ AMD Fusion ∎ Intel Sandy Bridge Page 30. Applied Mathematics Why bother? ...
Uploads
Papers by Geir Hasle