Pub Date : 2004-06-19DOI: 10.1109/CEC.2004.1330908
B. Filipič, T. Robic
In continuous casting of steel a number of parameters have to be set, such as the casting temperature, casting speed and coolant flows that critically affect the safety, quality and productivity of steel production. We have implemented an optimization tool consisting in an optimization algorithm and casting process simulator. The paper describes the process, the optimization task, and the proposed optimization approach, and shows illustrative results of its application on an industrial casting machine where spray coolant flows were optimized. In the comparative study, two variants of an evolutionary algorithm and the downhill simplex method were used, and they were all able to significantly improve the manual setting of coolant flows.
{"title":"A comparative study of coolant flow optimization on a steel casting machine","authors":"B. Filipič, T. Robic","doi":"10.1109/CEC.2004.1330908","DOIUrl":"https://doi.org/10.1109/CEC.2004.1330908","url":null,"abstract":"In continuous casting of steel a number of parameters have to be set, such as the casting temperature, casting speed and coolant flows that critically affect the safety, quality and productivity of steel production. We have implemented an optimization tool consisting in an optimization algorithm and casting process simulator. The paper describes the process, the optimization task, and the proposed optimization approach, and shows illustrative results of its application on an industrial casting machine where spray coolant flows were optimized. In the comparative study, two variants of an evolutionary algorithm and the downhill simplex method were used, and they were all able to significantly improve the manual setting of coolant flows.","PeriodicalId":152088,"journal":{"name":"Proceedings of the 2004 Congress on Evolutionary Computation (IEEE Cat. No.04TH8753)","volume":"38 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2004-06-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116650566","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 : 2004-06-19DOI: 10.1109/CEC.2004.1330898
Elizabeth León Guzman, O. Nasraoui, Jonatan Gómez
We present a new approach to anomaly detection based on unsupervised niche clustering (UNC). The UNC is a genetic niching technique for clustering that can handle noise, and is able to determine the number of clusters automatically. The UNC uses the normal samples for generating a profile of the normal space (clusters). Each cluster can later be characterized by a fuzzy membership function that follows a Gaussian shape defined by the evolved cluster centers and radii. The set of memberships are aggregated using a max-or fuzzy operator in order to determine the normalcy level of a data sample. Experiments on synthetic and real data sets, including a network intrusion detection data set, are performed and some results are analyzed and reported.
{"title":"Anomaly detection based on unsupervised niche clustering with application to network intrusion detection","authors":"Elizabeth León Guzman, O. Nasraoui, Jonatan Gómez","doi":"10.1109/CEC.2004.1330898","DOIUrl":"https://doi.org/10.1109/CEC.2004.1330898","url":null,"abstract":"We present a new approach to anomaly detection based on unsupervised niche clustering (UNC). The UNC is a genetic niching technique for clustering that can handle noise, and is able to determine the number of clusters automatically. The UNC uses the normal samples for generating a profile of the normal space (clusters). Each cluster can later be characterized by a fuzzy membership function that follows a Gaussian shape defined by the evolved cluster centers and radii. The set of memberships are aggregated using a max-or fuzzy operator in order to determine the normalcy level of a data sample. Experiments on synthetic and real data sets, including a network intrusion detection data set, are performed and some results are analyzed and reported.","PeriodicalId":152088,"journal":{"name":"Proceedings of the 2004 Congress on Evolutionary Computation (IEEE Cat. No.04TH8753)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2004-06-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121017616","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 : 2004-06-19DOI: 10.1109/CEC.2004.1330836
D. Arnold
Evolution strategies and gradient strategies are two different approaches to continuous optimization. Salomon's evolutionary gradient search procedure is a hybrid strategy that obtains gradient estimates by borrowing the idea of random variations from evolutionary computation. The present paper applies successful tools and ideas from the theory of evolution strategies to the evolutionary gradient search framework. Performance and the influence of its parameters. Comparisons with the (/spl mu///spl mu/,/spl lambda/)-ES are presented, and the issue of genetic repair in evolutionary gradient search is discussed. The practically relevant problem of noisy objective function measurements is addressed, and recommendations with regard to the setting of strategy parameters are made.
{"title":"An analysis of evolutionary gradient search","authors":"D. Arnold","doi":"10.1109/CEC.2004.1330836","DOIUrl":"https://doi.org/10.1109/CEC.2004.1330836","url":null,"abstract":"Evolution strategies and gradient strategies are two different approaches to continuous optimization. Salomon's evolutionary gradient search procedure is a hybrid strategy that obtains gradient estimates by borrowing the idea of random variations from evolutionary computation. The present paper applies successful tools and ideas from the theory of evolution strategies to the evolutionary gradient search framework. Performance and the influence of its parameters. Comparisons with the (/spl mu///spl mu/,/spl lambda/)-ES are presented, and the issue of genetic repair in evolutionary gradient search is discussed. The practically relevant problem of noisy objective function measurements is addressed, and recommendations with regard to the setting of strategy parameters are made.","PeriodicalId":152088,"journal":{"name":"Proceedings of the 2004 Congress on Evolutionary Computation (IEEE Cat. No.04TH8753)","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2004-06-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127381603","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 : 2004-06-19DOI: 10.1109/CEC.2004.1330986
T. Bartz-Beielstein, S. Markon
The optimization of complex real-world problems might benefit from well tuned algorithm's parameters. We propose a methodology that performs this tuning in an effective and efficient algorithmical manner. This approach combines methods from statistical design of experiments, regression analysis, design and analysis of computer experiments methods, and tree-based regression. It can also be applied to analyze the influence of different operators or to compare the performance of different algorithms. An evolution strategy and a simulated annealing algorithm that optimize an elevator supervisory group controller system are used to demonstrate the applicability of our approach to real-world optimization problems.
{"title":"Tuning search algorithms for real-world applications: a regression tree based approach","authors":"T. Bartz-Beielstein, S. Markon","doi":"10.1109/CEC.2004.1330986","DOIUrl":"https://doi.org/10.1109/CEC.2004.1330986","url":null,"abstract":"The optimization of complex real-world problems might benefit from well tuned algorithm's parameters. We propose a methodology that performs this tuning in an effective and efficient algorithmical manner. This approach combines methods from statistical design of experiments, regression analysis, design and analysis of computer experiments methods, and tree-based regression. It can also be applied to analyze the influence of different operators or to compare the performance of different algorithms. An evolution strategy and a simulated annealing algorithm that optimize an elevator supervisory group controller system are used to demonstrate the applicability of our approach to real-world optimization problems.","PeriodicalId":152088,"journal":{"name":"Proceedings of the 2004 Congress on Evolutionary Computation (IEEE Cat. No.04TH8753)","volume":"24 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2004-06-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125150128","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 : 2004-06-19DOI: 10.1109/CEC.2004.1330904
Yu Liu, Zheng Qin, Xingshi He
Particle swarm optimization (PSO) algorithms have exhibited good performance on well-known numerical test problems. In this paper, we propose a supervisor-student model in particle swarm optimization (SSM-PSO) that may further reduce computational cost in two aspects. On the one hand, it introduces a new parameter, called momentum factor, into the position update equation, which can restrict the particles inside the defined search space without checking the boundary at every iteration. On the other hand, relaxation-velocity-update strategy that is to update the velocities of the particles as few times as possible during the run, is employed to reduce the computational cost for evaluating the velocity. Comparisons with the linear decreasing weight PSO on three benchmark functions indicate that SSM-PSO not only greatly reduces the computational cost for updating the velocity, but also exhibit good performance.
{"title":"Supervisor-student model in particle swarm optimization","authors":"Yu Liu, Zheng Qin, Xingshi He","doi":"10.1109/CEC.2004.1330904","DOIUrl":"https://doi.org/10.1109/CEC.2004.1330904","url":null,"abstract":"Particle swarm optimization (PSO) algorithms have exhibited good performance on well-known numerical test problems. In this paper, we propose a supervisor-student model in particle swarm optimization (SSM-PSO) that may further reduce computational cost in two aspects. On the one hand, it introduces a new parameter, called momentum factor, into the position update equation, which can restrict the particles inside the defined search space without checking the boundary at every iteration. On the other hand, relaxation-velocity-update strategy that is to update the velocities of the particles as few times as possible during the run, is employed to reduce the computational cost for evaluating the velocity. Comparisons with the linear decreasing weight PSO on three benchmark functions indicate that SSM-PSO not only greatly reduces the computational cost for updating the velocity, but also exhibit good performance.","PeriodicalId":152088,"journal":{"name":"Proceedings of the 2004 Congress on Evolutionary Computation (IEEE Cat. No.04TH8753)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2004-06-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125898654","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 : 2004-06-19DOI: 10.1109/CEC.2004.1330895
Y. Shan, R. I. McKay, R. Baxter, H. Abbass, D. Essam, N. X. Hoai
In evolutionary computation, genetic operators, such as mutation and crossover, are employed to perturb individuals to generate the next population. However these fixed, problem independent genetic operators may destroy the sub-solution, usually called building blocks, instead of discovering and preserving them. One way to overcome this problem is to build a model based on the good individuals, and sample this model to obtain the next population. There is a wide range of such work in genetic algorithms; but because of the complexity of the genetic programming (GP) tree representation, little work of this kind has been done in GP. In this paper, we propose a new method, grammar model-based program evolution (GMPE) to evolved GP program. We replace common GP genetic operators with a probabilistic context-free grammar (SCFG). In each generation, an SCFG is learnt, and a new population is generated by sampling this SCFG model. On two benchmark problems we have studied, GMPE significantly outperforms conventional GP, learning faster and more reliably.
{"title":"Grammar model-based program evolution","authors":"Y. Shan, R. I. McKay, R. Baxter, H. Abbass, D. Essam, N. X. Hoai","doi":"10.1109/CEC.2004.1330895","DOIUrl":"https://doi.org/10.1109/CEC.2004.1330895","url":null,"abstract":"In evolutionary computation, genetic operators, such as mutation and crossover, are employed to perturb individuals to generate the next population. However these fixed, problem independent genetic operators may destroy the sub-solution, usually called building blocks, instead of discovering and preserving them. One way to overcome this problem is to build a model based on the good individuals, and sample this model to obtain the next population. There is a wide range of such work in genetic algorithms; but because of the complexity of the genetic programming (GP) tree representation, little work of this kind has been done in GP. In this paper, we propose a new method, grammar model-based program evolution (GMPE) to evolved GP program. We replace common GP genetic operators with a probabilistic context-free grammar (SCFG). In each generation, an SCFG is learnt, and a new population is generated by sampling this SCFG model. On two benchmark problems we have studied, GMPE significantly outperforms conventional GP, learning faster and more reliably.","PeriodicalId":152088,"journal":{"name":"Proceedings of the 2004 Congress on Evolutionary Computation (IEEE Cat. No.04TH8753)","volume":"82 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2004-06-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125935259","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 : 2004-06-19DOI: 10.1109/CEC.2004.1331110
C. Stephan, J. Sullivan
An agent-based model is presented of the transition of a personal transportation system based on conventional fuels to one based on an alternative fuel, such as hydrogen, requiring a new support infrastructure. The model allows two types of agents, vehicle owners and hydrogen fuel suppliers, to interact on a grid of roads representing a metropolitan region, and shows how their initial placement on the grid can lead either to successful or to unsuccessful transitions.
{"title":"An agent-based hydrogen vehicle/infrastructure model","authors":"C. Stephan, J. Sullivan","doi":"10.1109/CEC.2004.1331110","DOIUrl":"https://doi.org/10.1109/CEC.2004.1331110","url":null,"abstract":"An agent-based model is presented of the transition of a personal transportation system based on conventional fuels to one based on an alternative fuel, such as hydrogen, requiring a new support infrastructure. The model allows two types of agents, vehicle owners and hydrogen fuel suppliers, to interact on a grid of roads representing a metropolitan region, and shows how their initial placement on the grid can lead either to successful or to unsuccessful transitions.","PeriodicalId":152088,"journal":{"name":"Proceedings of the 2004 Congress on Evolutionary Computation (IEEE Cat. No.04TH8753)","volume":"58 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2004-06-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123306445","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 : 2004-06-19DOI: 10.1109/CEC.2004.1331075
J. Castro, André Seznec, P. I. Viñuela
The design of pseudorandom number generators by means of evolutionary computation is a classical problem. Today, it has been mostly and better accomplished by means of cellular automata and not many proposals, inside or outside this paradigm could claim to be both robust (passing all the statistical tests, including the most demanding ones) and fast, as is the case of the proposal we present here. Furthermore, for obtaining these generators, we use a radical approach, where our fitness function is not at all based in any measure of randomness, as is frequently the case in the literature, but of nonlinearity. Efficiency is assured by using only very efficient operators (both in hardware and software) and by limiting the number of terminals in the genetic programming implementation.
{"title":"On the design of state-of-the-art pseudorandom number generators by means of genetic programming","authors":"J. Castro, André Seznec, P. I. Viñuela","doi":"10.1109/CEC.2004.1331075","DOIUrl":"https://doi.org/10.1109/CEC.2004.1331075","url":null,"abstract":"The design of pseudorandom number generators by means of evolutionary computation is a classical problem. Today, it has been mostly and better accomplished by means of cellular automata and not many proposals, inside or outside this paradigm could claim to be both robust (passing all the statistical tests, including the most demanding ones) and fast, as is the case of the proposal we present here. Furthermore, for obtaining these generators, we use a radical approach, where our fitness function is not at all based in any measure of randomness, as is frequently the case in the literature, but of nonlinearity. Efficiency is assured by using only very efficient operators (both in hardware and software) and by limiting the number of terminals in the genetic programming implementation.","PeriodicalId":152088,"journal":{"name":"Proceedings of the 2004 Congress on Evolutionary Computation (IEEE Cat. No.04TH8753)","volume":"26 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2004-06-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114939363","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 : 2004-06-19DOI: 10.1109/CEC.2004.1331060
G. T. Pulido, C. Coello
This work presents a simple mechanism to handle constraints with a particle swarm optimization algorithm. Our proposal uses a simple criterion based on closeness of a particle to the feasible region in order to select a leader. Additionally, our algorithm incorporates a turbulence operator that improves the exploratory capabilities of our particle swarm optimization algorithm. Despite its relative simplicity, our comparison of results indicates that the proposed approach is highly competitive with respect to three constraint-handling techniques representative of the state-of-the-art in the area.
{"title":"A constraint-handling mechanism for particle swarm optimization","authors":"G. T. Pulido, C. Coello","doi":"10.1109/CEC.2004.1331060","DOIUrl":"https://doi.org/10.1109/CEC.2004.1331060","url":null,"abstract":"This work presents a simple mechanism to handle constraints with a particle swarm optimization algorithm. Our proposal uses a simple criterion based on closeness of a particle to the feasible region in order to select a leader. Additionally, our algorithm incorporates a turbulence operator that improves the exploratory capabilities of our particle swarm optimization algorithm. Despite its relative simplicity, our comparison of results indicates that the proposed approach is highly competitive with respect to three constraint-handling techniques representative of the state-of-the-art in the area.","PeriodicalId":152088,"journal":{"name":"Proceedings of the 2004 Congress on Evolutionary Computation (IEEE Cat. No.04TH8753)","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2004-06-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115198501","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 : 2004-06-19DOI: 10.1109/CEC.2004.1330838
Yoko Kobayashi, E. Aiyoshi
This paper deals with combinatorial optimization of permutation type using multi-agents algorithm (MAA). In order to improve optimization capability, we introduced the reinforcement learning and several processes into this MAA. Optimization capability of this algorithm was compared in traveling salesman problem and it provided better optimization results than the conventional MAA and genetic algorithm.
{"title":"Optimization algorithm using multi-agents and reinforcement learning","authors":"Yoko Kobayashi, E. Aiyoshi","doi":"10.1109/CEC.2004.1330838","DOIUrl":"https://doi.org/10.1109/CEC.2004.1330838","url":null,"abstract":"This paper deals with combinatorial optimization of permutation type using multi-agents algorithm (MAA). In order to improve optimization capability, we introduced the reinforcement learning and several processes into this MAA. Optimization capability of this algorithm was compared in traveling salesman problem and it provided better optimization results than the conventional MAA and genetic algorithm.","PeriodicalId":152088,"journal":{"name":"Proceedings of the 2004 Congress on Evolutionary Computation (IEEE Cat. No.04TH8753)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2004-06-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115544850","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：