Timo Knuutila
University of Turku, Technology Research Center, Faculty Member
We present a general framework for constructing characterizable inference algorithms for regular languages. This general approach is based on the fact, that certain families of non-trivial and infinite regular languages can be described... more
We present a general framework for constructing characterizable inference algorithms for regular languages. This general approach is based on the fact, that certain families of non-trivial and infinite regular languages can be described with finite tuples of finite sets of strings (or trees). It follows from this property, that the methods are able to inductively learn in the limit from
Research Interests:
Research Interests:
The world's longest established OR journal. An official journal of The OR Society. The Journal of the Operational Research Society is published 12 times a year by Palgrave Macmillan on behalf of the ORS. It is the aim of the Journal... more
The world's longest established OR journal. An official journal of The OR Society. The Journal of the Operational Research Society is published 12 times a year by Palgrave Macmillan on behalf of the ORS. It is the aim of the Journal to present papers which are relevant to practitioners ...
Research Interests: Information Systems, Marketing, Computer Science, Information Technology, Project Management, and 16 moreOperations Research, Production, Logistics, Control, Management Science, Forecasting, Scheduling, Operational Research, Optimization, Investment, Mathematical Sciences, Printed Circuit Board, Reliability, The, Location, and Communications Technology
Research Interests:
Research Interests:
Research Interests:
A method to detect and manage collision situations of the concurrent operations made by two robot arms in a shared work space is presented. Two different strategies, collision avoidance and in-turns operation are considered. The operation... more
A method to detect and manage collision situations of the concurrent operations made by two robot arms in a shared work space is presented. Two different strategies, collision avoidance and in-turns operation are considered. The operation of a dual-armed assembly robot is modeled using Petri nets. The model is then tested by the means of an animated virtual assembly system. By observing the operation of the virtual robot, the observer can verify that the robot arms work correctly and that the possible deadlock situations are avoided successfully. Finally, the performance of the collision avoidance method and that of the in-turn operation mode are compared.
Research Interests:
Collusion is covert co-operation between participants of a game. It poses a serious problem to multiplayer games that do not allow the players to share knowledge or resources with other players. Fair play requires that players act... more
Collusion is covert co-operation between participants of a game. It poses a serious problem to multiplayer games that do not allow the players to share knowledge or resources with other players. Fair play requires that players act independently to gain best possible individual result and are also able to analyse the action-reaction loop of the other players. In this paper, we discern two basic types of collusion - among the participants and among the players - and give a classification of different collusion types. Since collusion cannot be tackled using real-time methods, we propose how to counteract collusion using prevention and detection methods.
Research Interests:
The Printed Circuit Board (PCB) assembly business is a fast paced field of industry where the manufacturers must quickly adapt their production to meet the customer requirements. The goal of the production scheduling is to prioritize jobs... more
The Printed Circuit Board (PCB) assembly business is a fast paced field of industry where the manufacturers must quickly adapt their production to meet the customer requirements. The goal of the production scheduling is to prioritize jobs and optimize the usage of the production lines by allocating the jobs optimally to different lines. For the efficient usage of the production lines the workload balancing between the machines must be done properly. Balancing is usually a challenging operation and doing it consists of multiple calculations that solve the optimal placement time for each machine for different groups of components. In the present paper, the problem is modeled and solved using Mixed Integer Nonlinear Programming (MINLP) techniques. A pseudoconvex objective function for optimizing the production planning is presented. Different convexification techniques of non-linear functions are presented. The convexified model guarantees, in theory, that the global optimal solution w...
In this paper we address the combined problem of job-ordering and tool placement, where each tool can occupy more than one slot of the primary storage magazine. The capacity of the magazine is limited so that all the tools neccessary in... more
In this paper we address the combined problem of job-ordering and tool placement, where each tool can occupy more than one slot of the primary storage magazine. The capacity of the magazine is limited so that all the tools neccessary in the production cannot fit into the magazine at the same time, and the cost of magazine reorganization depends linearly
Research Interests:
Modular reconfigurable placement machines represent one of the most recent and most popular types of placement machines to respond to the needs for increased flexibility and productivity in automated printed circuit board (PCB) assembly.... more
Modular reconfigurable placement machines represent one of the most recent and most popular types of placement machines to respond to the needs for increased flexibility and productivity in automated printed circuit board (PCB) assembly. This paper studies the combined task of determining a favourable machine configuration and line balancing for an assembly line where a single type of PCB is
Research Interests:
Research Interests:
Research Interests:
ABSTRACT Emerging mobile application ecosystems have had a clear effect on the software business. Apple App Store and Google Play have gathered both existing large software companies and new start-ups. In creating a healthy ecosystem, the... more
ABSTRACT Emerging mobile application ecosystems have had a clear effect on the software business. Apple App Store and Google Play have gathered both existing large software companies and new start-ups. In creating a healthy ecosystem, the role of the software developer is significant. In practice, the ecosystems' ability to entice developers to create software products to the ecosystem can be argued to be a major factor driving the competitiveness of the ecosystem. This article empirically investigates Google Play, by gathering the data of 350,000 applications from the marketplace. With the dataset, common assumptions linked to the marketplace are studied. The results show that the direct software sale is a practical revenue model only for a few while offering a trial and paid version of the application seems to improve the revenue. The impact of the number of applications in the marketplace is questioned.
Research Interests:
ABSTRACT Software business has been changing from a customer-driven project business to a more concrete product business. This has also changed the conceptualizations related to software products and product development. Terms such as... more
ABSTRACT Software business has been changing from a customer-driven project business to a more concrete product business. This has also changed the conceptualizations related to software products and product development. Terms such as ‘modularity’, ‘standardization’ and ‘productization’ have been introduced to describe these changes. This study reviews different conceptualizations in creating software products and tries to identify the research gap that has been arisen from the change of the business. The study shows that the paradigm change to more market-driven products has introduced ambiguous terminology, which significantly overlaps with existing mature product development concepts. The study suggests that research should focus more on existing approaches in analyzing software products.
ABSTRACT Emerging mobile ecosystems have had a clear effect on the software product business. Marketplaces such as Apple App Store and Google Android Market have gathered both existing large software companies and new start-ups.... more
ABSTRACT Emerging mobile ecosystems have had a clear effect on the software product business. Marketplaces such as Apple App Store and Google Android Market have gathered both existing large software companies and new start-ups. Furthermore, the competition of the ecosystem providers will increase in the near future. In creating a healthy ecosystem, the role of the software developer is significant. In practise, the ecosystems‟ ability to entice developers to create software products to the ecosystem can be argued to be a major factor driving the competitiveness of the ecosystem. By offering a platform that enables the developers to easily create software products, which can be easily monetized, makes the platform more approachable. In this paper, we investigate an emerging mobile ecosystem, Android Market, its developers and users by gathering the data of 339,861 applications from the marketplace. With the dataset, we study some common assumptions linked to the marketplace. The results show that the direct software sale to the customers is a practical revenue model only for a few developers. The impact of the size of a marketplace is questioned by the results. However, offering a trial and paid version of the application seems to improve the revenue. Furthermore, certain conventional assumptions of users' behaviour do not seem to hold.
Research Interests:
Page 1. OR Spectrum (2008) 30:493513 DOI 10.1007/s00291-007-0108-1 REGULAR ARTICLE Selecting the nozzle assortment for a Gantry-type placement machine Csaba Raduly-Baka · Timo Knuutila · Mika Johnsson · Olli S. Nevalainen ...
Research Interests:
ABSTRACT Mobile application ecosystems have growth rapidly in the past few years. Increasing number of startups and established developers are alike offering their products in different marketplaces such as Android Market and Apple App... more
ABSTRACT Mobile application ecosystems have growth rapidly in the past few years. Increasing number of startups and established developers are alike offering their products in different marketplaces such as Android Market and Apple App Store. In this paper, we are studying revenue models used in Android Market. For analysis, we gathered the data of 351,601 applications from their public pages at the marketplace. From these, a random sample of 100 applications was used in a qualitative study of revenue streams. The results indicate that a part of the marketplace can be explained with traditional models but free applications use complex revenue models. Basing on the qualitative analysis, we identified four general business strategy categories for further studies.
Research Interests:
Research Interests:
Research Interests:
A method to detect and manage collision situations of the concurrent operations made by two robot arms in a shared work space is presented. Two different strategies, collision avoidance and in-turns operation are considered. The operation... more
A method to detect and manage collision situations of the concurrent operations made by two robot arms in a shared work space is presented. Two different strategies, collision avoidance and in-turns operation are considered. The operation of a dual-armed assembly robot is modeled using Petri nets. The model is then tested by the means of an animated virtual assembly system. By observing the operation of the virtual robot, the observer can verify that the robot arms work correctly and that the possible deadlock situations are avoided successfully. Finally, the performance of the collision avoidance method and that of the in-turn operation mode are compared.
Research Interests:
Several production planning tasks in the printed circuit board (PCB) assembly industry involve the estimation of the component placement times for different PCB types and placement machines. This kind of task may be, for example, the... more
Several production planning tasks in the printed circuit board (PCB) assembly industry involve the estimation of the component placement times for different PCB types and placement machines. This kind of task may be, for example, the scheduling of jobs or line balancing for single or multiple jobs. The simplest approach to time estimation is to let the production time be a linear function of the number of components to be placed. To achieve more accurate results, the model should include more parameters (e.g. the number of different component types, the number of different component shapes, the dimensions of the PCBs, etc.). In this study we train multilayer neural networks to approximate the assembly times of two different types of assembly machines based on several parameter combinations. It turns out that conventional learning methods are prone to overfitting when the number of hidden units of the network is large in relation to the number of training cases. To avoid this and complicated training and testing, we use Bayesian regularisation to achieve efficient learning and good accuracy automatically.
Research Interests:
The production efficiency of printed circuit board (PCB) assembly depends strongly on the organization of the component placement jobs. This is characteristic, especially in a high-mix low-volume production environment. The present study... more
The production efficiency of printed circuit board (PCB) assembly depends strongly on the organization of the component placement jobs. This is characteristic, especially in a high-mix low-volume production environment. The present study discusses the problem of arranging the jobs of ...
Research Interests:
ABSTRACT Production planning and control of the printed circuit board (PCB) assembly includes several decisions dealing with, for example, grouping of PCB jobs, allocation of PCB batches to machine lines, sequencing of batches and load... more
ABSTRACT Production planning and control of the printed circuit board (PCB) assembly includes several decisions dealing with, for example, grouping of PCB jobs, allocation of PCB batches to machine lines, sequencing of batches and load balancing of lines. The production time of a PCB job for a given placement machine is a key factor in this context and it must be quickly and accurately estimated, possibly millions of times in a single planning task, to avoid erroneous decisions. The commonly used nominal tact time-based estimators are very rough and the machine simulators too slow. Therefore, the purpose of this study is to give better machine-specific estimators that avoid the construction the actual machine control program. Two new estimators are proposed for gantry machines, one based on the information given by the manufacturer about the operations of the placement head, and the other on the regularised least-squares regression method trained with a set of PCB placement jobs. In practical evaluation with 95 PCB jobs, the mean absolute percentage error of the first and second methods are 3.75% and 6.52%, respectively, while that of the tact time-based approach is more than 17%. This indicates a great potential of the proposed methods as production time estimators.
Research Interests:
Ordering of batches of printed circuit boards (PCBs) has a significant impact on the efficiency of the electronic component placement processes. Through PCB batch grouping we aim to minimize the total setup time between batches. Batch... more
Ordering of batches of printed circuit boards (PCBs) has a significant impact on the efficiency of the electronic component placement processes. Through PCB batch grouping we aim to minimize the total setup time between batches. Batch groups are formed so that each group can be handled with one component setup. The job grouping problem calls for a set of groups
Research Interests:
Research Interests:
A variant of the classical job grouping problem (JGP) in printed circuit board (PCB) assembly is considered. Studies on JGPs have assumed a single feeder from which the components are retrieved and then placed on the PCB. Recent advances... more
A variant of the classical job grouping problem (JGP) in printed circuit board (PCB) assembly is considered. Studies on JGPs have assumed a single feeder from which the components are retrieved and then placed on the PCB. Recent advances in technology have made it ...
Research Interests:
Weighted grouping problems are shown to have an equivalent reduced form, which is often considerably smaller than the original problem. Although the reduction may be small for randomly generated problems, real-life problems often contain... more
Weighted grouping problems are shown to have an equivalent reduced form, which is often considerably smaller than the original problem. Although the reduction may be small for randomly generated problems, real-life problems often contain non-random properties that greatly increase the effect of reduction. We give an efficient algorithm to build the reduced problem instance, and analyse the expected amount of
Research Interests:
Research Interests:
This article considers a scheduling problem arising in flexible manufacturing systems. It is assumed that a computer numerical control machine processes a set of jobs with a set of wearing tools. The tool magazine of the machine has a... more
This article considers a scheduling problem arising in flexible manufacturing systems. It is assumed that a computer numerical control machine processes a set of jobs with a set of wearing tools. The tool magazine of the machine has a given capacity and each job requires some subset of tools. The goal is to minimize the average completion time of the
Research Interests:
Abstract Variable change costs between jobs in PCB assembly industry depend on the number of of component change occasions to be performed for the placement machine. Component changes are necessary due to the restricted feeder capacity of... more
Abstract Variable change costs between jobs in PCB assembly industry depend on the number of of component change occasions to be performed for the placement machine. Component changes are necessary due to the restricted feeder capacity of the machines. ...
Research Interests:
ABSTRACT In printed circuit board (PCB) manufacturing the revolver head gantry machines are nowadays popular because of their flexibility, accuracy and high speed. The operation control of this kind of machines includes, for example,... more
ABSTRACT In printed circuit board (PCB) manufacturing the revolver head gantry machines are nowadays popular because of their flexibility, accuracy and high speed. The operation control of this kind of machines includes, for example, selecting the nozzles into the revolver, assigning the component reels to feeder slots, and determining the pick-up and placement sequences of the components. We consider in this article the feeder assignment and the pick-up and placement sequencing problems.Unlike to some previous literature on these problems, we suppose that each component can be picked up by a nozzle of a certain type, only. For the feeder assignment problem a new heuristic is given and tested against four existing algorithms. The proposed heuristic considers the types of the r nearest neighbors of each component on the PCB and assigns component feeders close to each other according to the closeness of the component types on the PCB. The experimental tests are performed using two data sets based on realistic PCBs. The new heuristic outperformed the previous methods with 3.4% faster placement times. For determining the pick-up and placement sequences four rule-based algorithms are introduced and their performances evaluated. The best two of these construct the sequences greedily around the placement position of a starting component, which is selected in turns as the component nearest to each corner of the PCB.