Pub Date : 2018-07-01DOI: 10.1109/CEC.2018.8477902
V. S. Medeiros, A. Kubrusly, M. Jimenez, M. Freitas, J. P. Weid
Thickness measurement is a crucial matter in many applications, such as pipeline inspection by means of ultrasound. In this regard, an automated system must rely on an algorithm able to identify properly the ultrasonic echoes originated from the pipe's walls, which can be disturbed by noise and other sources of ultrasonic reflections. This paper describes the application of genetic algorithms for processing and analysis of signals obtained from thickness measurement using ultrasonic transducers. The main application for this algorithm is the processing and analysis of ultrasound signals obtained from oil duct inspections using ultrasonic pipeline inspection gauges (PIGs). The objective of the proposed algorithm is to identify correctly the ultrasound echoes in order to obtain an accurate thickness measurement, allowing the identification of cracks and corrosions in pipeline inspections. The algorithm was applied to several signals obtained from laboratory experiments with different distances between the transducer and a test plate with known thickness. Its efficiency was measured in terms of error percentage and computational cost.
{"title":"Application of Genetic Algorithms to Identify Ultrasonic Echoes for Thickness Measurements","authors":"V. S. Medeiros, A. Kubrusly, M. Jimenez, M. Freitas, J. P. Weid","doi":"10.1109/CEC.2018.8477902","DOIUrl":"https://doi.org/10.1109/CEC.2018.8477902","url":null,"abstract":"Thickness measurement is a crucial matter in many applications, such as pipeline inspection by means of ultrasound. In this regard, an automated system must rely on an algorithm able to identify properly the ultrasonic echoes originated from the pipe's walls, which can be disturbed by noise and other sources of ultrasonic reflections. This paper describes the application of genetic algorithms for processing and analysis of signals obtained from thickness measurement using ultrasonic transducers. The main application for this algorithm is the processing and analysis of ultrasound signals obtained from oil duct inspections using ultrasonic pipeline inspection gauges (PIGs). The objective of the proposed algorithm is to identify correctly the ultrasound echoes in order to obtain an accurate thickness measurement, allowing the identification of cracks and corrosions in pipeline inspections. The algorithm was applied to several signals obtained from laboratory experiments with different distances between the transducer and a test plate with known thickness. Its efficiency was measured in terms of error percentage and computational cost.","PeriodicalId":212677,"journal":{"name":"2018 IEEE Congress on Evolutionary Computation (CEC)","volume":"39 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128303874","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-07-01DOI: 10.1109/CEC.2018.8477815
Miriam Pescador-Rojas, C. Coello
In recent years, the use of decomposition-based multi-objective evolutionary algorithms has been very successful in solving both multi- and many-objective optimization problems. In these algorithms, the adopted Scalarizing Functions (SFs) play a crucial role in their performance. Methods such as the Modified Weighted Chebyshev (MCHE), Penalty Boundary Intersection (PBI) and Augmented Achievement Scalarizing Function (AASF) have been found to be very effective for achieving both convergence to the true Pareto front and a uniform distribution of solutions along it. However, the choice of an appropriate model parameter is required for these SFs. Some studies have analyzed the impact of these parameter values on the performance of the best-known decomposition multi-objective evolutionary algorithm (MOEA/D). In this paper, we propose a strategy based on collaborative populations combining different SFs and model parameter values via an adaptive operator selection based on the multi-armed bandit technique. Our preliminary results give rise to some interesting observations regarding the way in which different SFs are combined and adapted during the evolutionary process of MOEA/D.
{"title":"Collaborative and Adaptive Strategies of Different Scalarizing Functions in MOEA/D","authors":"Miriam Pescador-Rojas, C. Coello","doi":"10.1109/CEC.2018.8477815","DOIUrl":"https://doi.org/10.1109/CEC.2018.8477815","url":null,"abstract":"In recent years, the use of decomposition-based multi-objective evolutionary algorithms has been very successful in solving both multi- and many-objective optimization problems. In these algorithms, the adopted Scalarizing Functions (SFs) play a crucial role in their performance. Methods such as the Modified Weighted Chebyshev (MCHE), Penalty Boundary Intersection (PBI) and Augmented Achievement Scalarizing Function (AASF) have been found to be very effective for achieving both convergence to the true Pareto front and a uniform distribution of solutions along it. However, the choice of an appropriate model parameter is required for these SFs. Some studies have analyzed the impact of these parameter values on the performance of the best-known decomposition multi-objective evolutionary algorithm (MOEA/D). In this paper, we propose a strategy based on collaborative populations combining different SFs and model parameter values via an adaptive operator selection based on the multi-armed bandit technique. Our preliminary results give rise to some interesting observations regarding the way in which different SFs are combined and adapted during the evolutionary process of MOEA/D.","PeriodicalId":212677,"journal":{"name":"2018 IEEE Congress on Evolutionary Computation (CEC)","volume":"43 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128330218","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-07-01DOI: 10.1109/CEC.2018.8477888
R. P. C. Moreira, E. Wanner, F. V. Martins, J. Sarubbi
This work proposes an evolutionary approach to solve the Menu Planning Problem. Our work uses the Brazilian school context and our principal goal is to create menus that minimize the total cost of these menus. However, those menus must also satisfy requirements of the Brazilian government, such as: (i) student age group, (ii) school category, (iii) school duration time, (iv) school location, (v) variety of preparations, (vi) harmony of preparations, (vii) maximum amount to be paid for each meal and, (viii) lower and upper limits of macronutrients. The results demonstrate that the evolutionary approach is not only able to generate a set of inexpensive and healthy menus but also respect the required set of constraints. A constrained deterministic approach is performed to generate 5-day menu through a greedy-based function taking into account the normalized sum of all macronutrients and the monetary cost of the menu. A comparison between the 5-day menu obtained by the proposed approach and the constrained greedy-based approach menu is carried out. Despite the fact the obtained menu outperforms the greed-based menu taking into account the total cost, this difference is not so expressive. However, all macronutrients were outside the pre-defined range at least in one day of the week. The 5-day menu obtained by the proposed approach is evaluated by a nutritionist. The overall quality of the menu is outstanding and the time spent to generate it is 60 seconds.
{"title":"An Evolutionary Mono-Objective Approach for Solving the Menu Planning Problem","authors":"R. P. C. Moreira, E. Wanner, F. V. Martins, J. Sarubbi","doi":"10.1109/CEC.2018.8477888","DOIUrl":"https://doi.org/10.1109/CEC.2018.8477888","url":null,"abstract":"This work proposes an evolutionary approach to solve the Menu Planning Problem. Our work uses the Brazilian school context and our principal goal is to create menus that minimize the total cost of these menus. However, those menus must also satisfy requirements of the Brazilian government, such as: (i) student age group, (ii) school category, (iii) school duration time, (iv) school location, (v) variety of preparations, (vi) harmony of preparations, (vii) maximum amount to be paid for each meal and, (viii) lower and upper limits of macronutrients. The results demonstrate that the evolutionary approach is not only able to generate a set of inexpensive and healthy menus but also respect the required set of constraints. A constrained deterministic approach is performed to generate 5-day menu through a greedy-based function taking into account the normalized sum of all macronutrients and the monetary cost of the menu. A comparison between the 5-day menu obtained by the proposed approach and the constrained greedy-based approach menu is carried out. Despite the fact the obtained menu outperforms the greed-based menu taking into account the total cost, this difference is not so expressive. However, all macronutrients were outside the pre-defined range at least in one day of the week. The 5-day menu obtained by the proposed approach is evaluated by a nutritionist. The overall quality of the menu is outstanding and the time spent to generate it is 60 seconds.","PeriodicalId":212677,"journal":{"name":"2018 IEEE Congress on Evolutionary Computation (CEC)","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128554743","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-07-01DOI: 10.1109/CEC.2018.8477905
William Higino, A. A. Chaves, V. V. D. Melo
Among the different classes of Vehicle Routing Problems are the Vehicle Routing Problems with Profits (VRPPs), where it is not mandatory to service all the customers. A relatively new VRPP is the VRPPFCC (Vehicle Routing Problem with Private Fleet and Common Carrier). In this problem, it is sometimes useful to directly serve only part of the shipping demand, outsourcing the rest of it to other companies. This paper presents the combination between the Biased Random-key Genetic Algorithm (BRKGA) and Random Variable Neighborhood Descent (RVND), a local search procedure, in the solution of the VRPPFCC. The implementation uses a vector of random keys as solution representation; thus a decoding heuristic is also developed, converting random keys to actual solutions for the VRPPFCC. Computational tests and conclusions focus on the comparison of the effectiveness of the methods, comparing their obtained solutions to the best known solutions for the problem.
{"title":"Biased Random-Key Genetic Algorithm Applied to the Vehicle Routing Problem with Private Fleet and Common Carrier","authors":"William Higino, A. A. Chaves, V. V. D. Melo","doi":"10.1109/CEC.2018.8477905","DOIUrl":"https://doi.org/10.1109/CEC.2018.8477905","url":null,"abstract":"Among the different classes of Vehicle Routing Problems are the Vehicle Routing Problems with Profits (VRPPs), where it is not mandatory to service all the customers. A relatively new VRPP is the VRPPFCC (Vehicle Routing Problem with Private Fleet and Common Carrier). In this problem, it is sometimes useful to directly serve only part of the shipping demand, outsourcing the rest of it to other companies. This paper presents the combination between the Biased Random-key Genetic Algorithm (BRKGA) and Random Variable Neighborhood Descent (RVND), a local search procedure, in the solution of the VRPPFCC. The implementation uses a vector of random keys as solution representation; thus a decoding heuristic is also developed, converting random keys to actual solutions for the VRPPFCC. Computational tests and conclusions focus on the comparison of the effectiveness of the methods, comparing their obtained solutions to the best known solutions for the problem.","PeriodicalId":212677,"journal":{"name":"2018 IEEE Congress on Evolutionary Computation (CEC)","volume":"38 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114614886","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-07-01DOI: 10.1109/CEC.2018.8477655
Rohit Salgotra, Urvinder Singh, S. Saha
Cuckoo Search is a nature inspired evolutionary algorithm to solve real-world optimization problems. It is inspired from the brood parasitism of cuckoos. It is highly competitive and has been used to solve number of problems in the field of science and engineering. A number of modifications have been proposed to enhance its performance in the past. This paper presents an improved version of CS namely CVnew in which three modifications are proposed. The first modification is the introduction of two new search equations to improve the global search while the second one deals with the incorporation of four search equations to improve the local search. As a third modification, a balance between global and local search has been increased by exponentially decreasing the switch probability. The proposed algorithm has been applied to solve single objective real-parameter problems of CEC 2017. The numerical results prove the better performance of CVnew in comparison with SaDE, JADE, SHADE and MVMO.
{"title":"Improved Cuckoo Search with Better Search Capabilities for Solving CEC2017 Benchmark Problems","authors":"Rohit Salgotra, Urvinder Singh, S. Saha","doi":"10.1109/CEC.2018.8477655","DOIUrl":"https://doi.org/10.1109/CEC.2018.8477655","url":null,"abstract":"Cuckoo Search is a nature inspired evolutionary algorithm to solve real-world optimization problems. It is inspired from the brood parasitism of cuckoos. It is highly competitive and has been used to solve number of problems in the field of science and engineering. A number of modifications have been proposed to enhance its performance in the past. This paper presents an improved version of CS namely CVnew in which three modifications are proposed. The first modification is the introduction of two new search equations to improve the global search while the second one deals with the incorporation of four search equations to improve the local search. As a third modification, a balance between global and local search has been increased by exponentially decreasing the switch probability. The proposed algorithm has been applied to solve single objective real-parameter problems of CEC 2017. The numerical results prove the better performance of CVnew in comparison with SaDE, JADE, SHADE and MVMO.","PeriodicalId":212677,"journal":{"name":"2018 IEEE Congress on Evolutionary Computation (CEC)","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121639174","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-07-01DOI: 10.1109/CEC.2018.8477922
Mathew Curtis, A. Lewis
The canonical MOPSO algorithm was adapted to include behaviour observed in starling flocks. An observer can see the amazing aerial display of large starling flocks. They maintain uniformity and cohesion throughout flight and landing. This behaviour emerges from the individuals following a set of simple rules governing motion and interaction. The adaption to the canonical MOPSO was done by extracting these rules to provide the algorithm with behaviour that improved uniformity and spread of the archived solutions. The adapted MOPSO was applied to ZDT1 - ZDT4. There was significant improvement in uniformity and spreading of the final archive solutions. The improvement in coverage was as high as 25.4% in the case of ZDT4. There was also an improvement in spread: ZDT1 by a factor of 8.4, ZDT2 by a factor of 4.78, ZDT3 by a factor of 1.6, and ZDT4 by a factor of 3.76. Local search was then added to the algorithm. The convergence showed significant improvement without loss of the newly improved coverage and spread. With better understanding of how and why behaviour emerges, we were able to improve the canonical MOPSO by adapting its fundamental rules leading to emergent behaviour that intrinsically improved deficiencies in uniformity and spread of archive solutions.
{"title":"Back to nature: improving MOPSO inspired by the behaviour of starlings","authors":"Mathew Curtis, A. Lewis","doi":"10.1109/CEC.2018.8477922","DOIUrl":"https://doi.org/10.1109/CEC.2018.8477922","url":null,"abstract":"The canonical MOPSO algorithm was adapted to include behaviour observed in starling flocks. An observer can see the amazing aerial display of large starling flocks. They maintain uniformity and cohesion throughout flight and landing. This behaviour emerges from the individuals following a set of simple rules governing motion and interaction. The adaption to the canonical MOPSO was done by extracting these rules to provide the algorithm with behaviour that improved uniformity and spread of the archived solutions. The adapted MOPSO was applied to ZDT1 - ZDT4. There was significant improvement in uniformity and spreading of the final archive solutions. The improvement in coverage was as high as 25.4% in the case of ZDT4. There was also an improvement in spread: ZDT1 by a factor of 8.4, ZDT2 by a factor of 4.78, ZDT3 by a factor of 1.6, and ZDT4 by a factor of 3.76. Local search was then added to the algorithm. The convergence showed significant improvement without loss of the newly improved coverage and spread. With better understanding of how and why behaviour emerges, we were able to improve the canonical MOPSO by adapting its fundamental rules leading to emergent behaviour that intrinsically improved deficiencies in uniformity and spread of archive solutions.","PeriodicalId":212677,"journal":{"name":"2018 IEEE Congress on Evolutionary Computation (CEC)","volume":"55 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130198927","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-07-01DOI: 10.1109/CEC.2018.8477927
Eduardo Corrêa Gonçalves, A. Freitas, A. Plastino
In recent years, multi-label classification (MLC) has become an emerging research topic in big data analytics and machine learning. In this problem, each object of a dataset may belong to multiple class labels and the goal is to learn a classification model that can infer the correct labels of new, previously unseen, objects. This paper presents a survey of genetic algorithms (GAs) designed for MLC tasks. The study is organized in three parts. First, we propose a new taxonomy focused on GAs for MLC. In the second part, we provide an up-to-date overview of the work in this area, categorizing the approaches identified in the literature with respect to the taxonomy. In the third and last part, we discuss some new ideas for combining GAs with MLC.
{"title":"A Survey of Genetic Algorithms for Multi-Label Classification","authors":"Eduardo Corrêa Gonçalves, A. Freitas, A. Plastino","doi":"10.1109/CEC.2018.8477927","DOIUrl":"https://doi.org/10.1109/CEC.2018.8477927","url":null,"abstract":"In recent years, multi-label classification (MLC) has become an emerging research topic in big data analytics and machine learning. In this problem, each object of a dataset may belong to multiple class labels and the goal is to learn a classification model that can infer the correct labels of new, previously unseen, objects. This paper presents a survey of genetic algorithms (GAs) designed for MLC tasks. The study is organized in three parts. First, we propose a new taxonomy focused on GAs for MLC. In the second part, we provide an up-to-date overview of the work in this area, categorizing the approaches identified in the literature with respect to the taxonomy. In the third and last part, we discuss some new ideas for combining GAs with MLC.","PeriodicalId":212677,"journal":{"name":"2018 IEEE Congress on Evolutionary Computation (CEC)","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133989151","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-07-01DOI: 10.1109/CEC.2018.8477677
S. Picek, Karlo Knezevic, L. Mariot, D. Jakobović, A. Leporati
Boolean functions have a prominent role in many real-world applications, which makes them a very active research domain. Throughout the years, various heuristic techniques proved to be an attractive choice for the construction of Boolean functions with different properties. One of the most important properties is nonlinearity, and in particular maximally nonlinear Boolean functions are also called bent functions. In this paper, instead of considering Boolean functions, we experiment with quaternary functions. The corresponding problem is much more difficult and presents an interesting benchmark as well as realworld applications. The results we obtain show that evolutionary metaheuristics, especially genetic programming, succeed in finding quaternary functions with the desired properties. The obtained results in the quaternary domain can also be translated into the binary domain, in which case this approach compares favorably with the state-of-the-art in Boolean optimization. Our techniques are able to find quaternary bent functions for up to 8 inputs, which corresponds to obtaining Boolean bent functions of 16 inputs.
{"title":"Evolving Bent Quaternary Functions","authors":"S. Picek, Karlo Knezevic, L. Mariot, D. Jakobović, A. Leporati","doi":"10.1109/CEC.2018.8477677","DOIUrl":"https://doi.org/10.1109/CEC.2018.8477677","url":null,"abstract":"Boolean functions have a prominent role in many real-world applications, which makes them a very active research domain. Throughout the years, various heuristic techniques proved to be an attractive choice for the construction of Boolean functions with different properties. One of the most important properties is nonlinearity, and in particular maximally nonlinear Boolean functions are also called bent functions. In this paper, instead of considering Boolean functions, we experiment with quaternary functions. The corresponding problem is much more difficult and presents an interesting benchmark as well as realworld applications. The results we obtain show that evolutionary metaheuristics, especially genetic programming, succeed in finding quaternary functions with the desired properties. The obtained results in the quaternary domain can also be translated into the binary domain, in which case this approach compares favorably with the state-of-the-art in Boolean optimization. Our techniques are able to find quaternary bent functions for up to 8 inputs, which corresponds to obtaining Boolean bent functions of 16 inputs.","PeriodicalId":212677,"journal":{"name":"2018 IEEE Congress on Evolutionary Computation (CEC)","volume":"321 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134050808","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-07-01DOI: 10.1109/CEC.2018.8477792
M. Lacerda, H. A. A. Neto, Teresa B Ludermir, H. Kuchen, Fernando Buarque de Lima-Neto
This paper proposes a mechanism of dynamic adjustment of the population size of population based metaheuristics in order to balance its efficacy and efficiency. In this approach, an external trajectory based metaheuristic (MH) is used to dynamically adjust the population size of an inner population based metaheuristic. A Particle Swarm Optmization (PSO) implemented for a Compute Unified Device Architecture platform (CUDA), called CUDA-PSO, is used as inner MH, while a sequential Simulated Annealing (SA) is used as an external one. The main objective of this paper is to evaluate the SA capabilities of finding a good balance between efficiency and efficacy during the CUDA-PSO execution and to assess its adaptability to different hardwares without any prior information about the computing platform. The results show that the new approach was able to find a good balance in most cases. Also, it was observed that this approach is able to adapt its operation to different hardwares.
{"title":"Population Size Control for Efficiency and Efficacy Optimization in Population Based Metaheuristics","authors":"M. Lacerda, H. A. A. Neto, Teresa B Ludermir, H. Kuchen, Fernando Buarque de Lima-Neto","doi":"10.1109/CEC.2018.8477792","DOIUrl":"https://doi.org/10.1109/CEC.2018.8477792","url":null,"abstract":"This paper proposes a mechanism of dynamic adjustment of the population size of population based metaheuristics in order to balance its efficacy and efficiency. In this approach, an external trajectory based metaheuristic (MH) is used to dynamically adjust the population size of an inner population based metaheuristic. A Particle Swarm Optmization (PSO) implemented for a Compute Unified Device Architecture platform (CUDA), called CUDA-PSO, is used as inner MH, while a sequential Simulated Annealing (SA) is used as an external one. The main objective of this paper is to evaluate the SA capabilities of finding a good balance between efficiency and efficacy during the CUDA-PSO execution and to assess its adaptability to different hardwares without any prior information about the computing platform. The results show that the new approach was able to find a good balance in most cases. Also, it was observed that this approach is able to adapt its operation to different hardwares.","PeriodicalId":212677,"journal":{"name":"2018 IEEE Congress on Evolutionary Computation (CEC)","volume":"49 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134364073","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-07-01DOI: 10.1109/CEC.2018.8477887
Aritz Pérez Martínez, Josu Ceberio, J. A. Lozano
In the field of evolutionary computation, it is usual to generate artificial benchmarks of instances that are used as a test-bed to determine the performance of the algorithms at hand. In this context, a recent work on permutation problems analyzed the implications of generating instances uniformly at random (u.a.r.) when building those benchmarks. Particularly, the authors analyzed instances as rankings of the solutions of the search space sorted according to their objective function value. Thus, two instances are considered equivalent when their objective functions induce the same ranking over the search space. Based on the analysis, they suggested that, when some restrictions hold, the probability to create easy rankings is higher than creating difficult ones. In this paper, we continue on that research line by adopting the framework of local search algorithms with the best improvement criterion. Particularly, we empirically analyze, in terms of difficulty, the instances (rankings) created u.a.r. of three popular problems: Linear Ordering Problem, Quadratic Assignment Problem and Flowshop Scheduling Problem. As the neighborhood system is critical for the performance of local search algorithms three different neighborhood systems have been considered: swap, interchange and insert. Conducted experiments reveal that (1) by sampling the parameters uniformly at random we obtain instances with a non-uniform distribution in terms of difficulty, (2) the distribution of the difficulty strongly depends on the pair problem-neighborhood considered, and (3) given a problem, the distribution of the difficulty seems to depend on the smoothness of the landscape induced by the neighborhood and on its size.
{"title":"Are the Artificially Generated Instances Uniform in Terms of Difficulty?","authors":"Aritz Pérez Martínez, Josu Ceberio, J. A. Lozano","doi":"10.1109/CEC.2018.8477887","DOIUrl":"https://doi.org/10.1109/CEC.2018.8477887","url":null,"abstract":"In the field of evolutionary computation, it is usual to generate artificial benchmarks of instances that are used as a test-bed to determine the performance of the algorithms at hand. In this context, a recent work on permutation problems analyzed the implications of generating instances uniformly at random (u.a.r.) when building those benchmarks. Particularly, the authors analyzed instances as rankings of the solutions of the search space sorted according to their objective function value. Thus, two instances are considered equivalent when their objective functions induce the same ranking over the search space. Based on the analysis, they suggested that, when some restrictions hold, the probability to create easy rankings is higher than creating difficult ones. In this paper, we continue on that research line by adopting the framework of local search algorithms with the best improvement criterion. Particularly, we empirically analyze, in terms of difficulty, the instances (rankings) created u.a.r. of three popular problems: Linear Ordering Problem, Quadratic Assignment Problem and Flowshop Scheduling Problem. As the neighborhood system is critical for the performance of local search algorithms three different neighborhood systems have been considered: swap, interchange and insert. Conducted experiments reveal that (1) by sampling the parameters uniformly at random we obtain instances with a non-uniform distribution in terms of difficulty, (2) the distribution of the difficulty strongly depends on the pair problem-neighborhood considered, and (3) given a problem, the distribution of the difficulty seems to depend on the smoothness of the landscape induced by the neighborhood and on its size.","PeriodicalId":212677,"journal":{"name":"2018 IEEE Congress on Evolutionary Computation (CEC)","volume":"24 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132443605","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：