S. Salcedo-Sanz, E. G. Ortíz-García, Ángel M. Pérez-Bellido, J. A. Portilla-Figueras, X. Yao
This paper presents two heuristics algorithms to solve Japanese puzzles, both black and white puzzles and color puzzles. First, we present ad-hoc heuristics which use the information in rows, columns, and puzzle's constraints to obtain the solution of the puzzle. The best heuristic developed for black and white puzzles is then extended to solving color Japanese puzzles. We show the performance of the proposed heuristics in several examples from a well known Web page devoted to this kind of puzzles. Comparison with an existing solver based on constraint programming and with a genetic algorithm is carried out
{"title":"Solving Japanese Puzzles with Heuristics","authors":"S. Salcedo-Sanz, E. G. Ortíz-García, Ángel M. Pérez-Bellido, J. A. Portilla-Figueras, X. Yao","doi":"10.1109/CIG.2007.368102","DOIUrl":"https://doi.org/10.1109/CIG.2007.368102","url":null,"abstract":"This paper presents two heuristics algorithms to solve Japanese puzzles, both black and white puzzles and color puzzles. First, we present ad-hoc heuristics which use the information in rows, columns, and puzzle's constraints to obtain the solution of the puzzle. The best heuristic developed for black and white puzzles is then extended to solving color Japanese puzzles. We show the performance of the proposed heuristics in several examples from a well known Web page devoted to this kind of puzzles. Comparison with an existing solver based on constraint programming and with a genetic algorithm is carried out","PeriodicalId":365269,"journal":{"name":"2007 IEEE Symposium on Computational Intelligence and Games","volume":"76 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127467254","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}
This paper considers variations on an extremely simple form of car racing, the challenge being to visit as many way-points as possible in a fixed amount of time. The simplicity of the models enables a very thorough evaluation of various learning algorithms and control architectures, and enables other researchers to work on the same models with relative ease. The models are used to compare the performance of various hand-programmed controllers, and neural networks trained using evolution, and using temporal difference learning. Comparisons are also made between state-based and action-based controller architectures. The best controllers were obtained using evolution to learn the weights of state-evaluation neural networks, and these were greatly superior to human drivers
{"title":"Point-to-Point Car Racing: an Initial Study of Evolution Versus Temporal Difference Learning","authors":"S. Lucas, J. Togelius","doi":"10.1109/CIG.2007.368107","DOIUrl":"https://doi.org/10.1109/CIG.2007.368107","url":null,"abstract":"This paper considers variations on an extremely simple form of car racing, the challenge being to visit as many way-points as possible in a fixed amount of time. The simplicity of the models enables a very thorough evaluation of various learning algorithms and control architectures, and enables other researchers to work on the same models with relative ease. The models are used to compare the performance of various hand-programmed controllers, and neural networks trained using evolution, and using temporal difference learning. Comparisons are also made between state-based and action-based controller architectures. The best controllers were obtained using evolution to learn the weights of state-evaluation neural networks, and these were greatly superior to human drivers","PeriodicalId":365269,"journal":{"name":"2007 IEEE Symposium on Computational Intelligence and Games","volume":"70 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131956227","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}
This paper introduces tournament particle swarm optimization (PSO) as a method to optimize weights of game tree evaluation functions in a competitive environment using particle swarm optimization. This method makes use of tournaments to ensure a fair evaluation of the performance of particles in the swarm, relative to that of other particles. The empirical work presented compares the performance of different tournament methods that can be applied to the tournament PSO, with application to Checkers.
{"title":"Tournament Particle Swarm Optimization","authors":"W. H. Duminy, A. Engelbrecht","doi":"10.1109/CIG.2007.368091","DOIUrl":"https://doi.org/10.1109/CIG.2007.368091","url":null,"abstract":"This paper introduces tournament particle swarm optimization (PSO) as a method to optimize weights of game tree evaluation functions in a competitive environment using particle swarm optimization. This method makes use of tournaments to ensure a fair evaluation of the performance of particles in the swarm, relative to that of other particles. The empirical work presented compares the performance of different tournament methods that can be applied to the tournament PSO, with application to Checkers.","PeriodicalId":365269,"journal":{"name":"2007 IEEE Symposium on Computational Intelligence and Games","volume":"49 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114700378","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}
Transposition tables are one common method to improve an alpha-beta searcher. We present two methods for extending the usage of transposition tables to chance nodes during stochastic game tree search. Empirical results show that these techniques can reduce the search effort of Ballard's Star2 algorithm by 37 percent.
{"title":"Effective Use of Transposition Tables in Stochastic Game Tree Search","authors":"J. Veness, Alan Blair","doi":"10.1109/CIG.2007.368086","DOIUrl":"https://doi.org/10.1109/CIG.2007.368086","url":null,"abstract":"Transposition tables are one common method to improve an alpha-beta searcher. We present two methods for extending the usage of transposition tables to chance nodes during stochastic game tree search. Empirical results show that these techniques can reduce the search effort of Ballard's Star2 algorithm by 37 percent.","PeriodicalId":365269,"journal":{"name":"2007 IEEE Symposium on Computational Intelligence and Games","volume":"35 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121538861","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}
Coevolutionary techniques have been proven to be effective in evolving solutions to many game related problems, with successful applications in many complex chess-like games like Othello, Checkers and Western Chess. This paper explores the application of coevolutionary models to learn Chinese Chess strategies. The proposed Chinese Chess engine uses alpha-beta search algorithm, quiescence search and move ordering. Three different models are studied: single-population competitive, host-parasite competitive and cooperative coevolutionary models. A modified alpha-beta algorithm is also developed for performance evaluation and an archiving mechanism is implemented to handle intransitive behaviour. Interesting traits are revealed when the coevolution models are simulated under different settings - with and without opening book. Results show that the coevolved players can perform relatively well, with the cooperative model being best for finding good players under random strategy initialization and the host-parasite model being best for the case when strategies are initialized with a good set of starting seeds.
{"title":"Discovering Chinese Chess Strategies through Coevolutionary Approaches","authors":"Chin Soon Ong, H. Quek, K. Tan, A. Tay","doi":"10.1109/CIG.2007.368121","DOIUrl":"https://doi.org/10.1109/CIG.2007.368121","url":null,"abstract":"Coevolutionary techniques have been proven to be effective in evolving solutions to many game related problems, with successful applications in many complex chess-like games like Othello, Checkers and Western Chess. This paper explores the application of coevolutionary models to learn Chinese Chess strategies. The proposed Chinese Chess engine uses alpha-beta search algorithm, quiescence search and move ordering. Three different models are studied: single-population competitive, host-parasite competitive and cooperative coevolutionary models. A modified alpha-beta algorithm is also developed for performance evaluation and an archiving mechanism is implemented to handle intransitive behaviour. Interesting traits are revealed when the coevolution models are simulated under different settings - with and without opening book. Results show that the coevolved players can perform relatively well, with the cooperative model being best for finding good players under random strategy initialization and the host-parasite model being best for the case when strategies are initialized with a good set of starting seeds.","PeriodicalId":365269,"journal":{"name":"2007 IEEE Symposium on Computational Intelligence and Games","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127650129","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}
In this paper we present a new method for evolving autonomous agents that are competitive in the space combat game Xpilot. A genetic algorithm is used to evolve the parameters related to the sensitivity of the agent to input stimuli and the agent's level of reaction to these stimuli. The resultant controllers are comparable to the best hand programmed artificial Xpilot bots, are competitive with human players, and display interesting behaviors that resemble human strategies.
{"title":"Evolving Parameters for Xpilot Combat Agents","authors":"G. Parker, M. Parker","doi":"10.1109/CIG.2007.368104","DOIUrl":"https://doi.org/10.1109/CIG.2007.368104","url":null,"abstract":"In this paper we present a new method for evolving autonomous agents that are competitive in the space combat game Xpilot. A genetic algorithm is used to evolve the parameters related to the sensitivity of the agent to input stimuli and the agent's level of reaction to these stimuli. The resultant controllers are comparable to the best hand programmed artificial Xpilot bots, are competitive with human players, and display interesting behaviors that resemble human strategies.","PeriodicalId":365269,"journal":{"name":"2007 IEEE Symposium on Computational Intelligence and Games","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129219458","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}
In this paper we address the problem of cooperation and selfish behavior in ad hoc networks. We present a new game theory based model to study cooperation between nodes. This model has some similarities with the iterated prisoner's dilemma under the random pairing game. In such game randomly chosen players receive payoffs that depend on the way they behave. The network gaming model includes a simple reputation collection and trust evaluation mechanisms. In our proposition a decision whether to forward or discard a packet is determined by a strategy based on the trust level in the source node of the packet and some general information about behavior of the network. A genetic algorithm (GA) is applied to evolve strategies for the participating nodes. These strategies are targeted to maximize the throughput of the network by enforcing cooperation. Experimental results show that proposed strategy based approach successfully enforces cooperation maximizing the network throughput
{"title":"Modelling the Evolution of Cooperative Behavior in Ad Hoc Networks using a Game Based Model","authors":"M. Seredynski, P. Bouvry, M. Kłopotek","doi":"10.1109/CIG.2007.368084","DOIUrl":"https://doi.org/10.1109/CIG.2007.368084","url":null,"abstract":"In this paper we address the problem of cooperation and selfish behavior in ad hoc networks. We present a new game theory based model to study cooperation between nodes. This model has some similarities with the iterated prisoner's dilemma under the random pairing game. In such game randomly chosen players receive payoffs that depend on the way they behave. The network gaming model includes a simple reputation collection and trust evaluation mechanisms. In our proposition a decision whether to forward or discard a packet is determined by a strategy based on the trust level in the source node of the packet and some general information about behavior of the network. A genetic algorithm (GA) is applied to evolve strategies for the participating nodes. These strategies are targeted to maximize the throughput of the network by enforcing cooperation. Experimental results show that proposed strategy based approach successfully enforces cooperation maximizing the network throughput","PeriodicalId":365269,"journal":{"name":"2007 IEEE Symposium on Computational Intelligence and Games","volume":"219 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132028051","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}
We address the problem of predicting moves in the board game of Go. We use the relative frequencies of local board patterns observed in game records to generate a ranked list of moves, and then apply the maximum entropy method (MEM) to the list to re-rank the moves. Move prediction is the task of selecting a small number of promising moves from all legal moves, and move prediction output can be used to improve the efficiency of the game tree search. The MEM enables us to make use of multiple overlapping features, while avoiding problems with data sparseness. Our system was trained on 20000 expert games and had 33.9% prediction accuracy in 500 expert games
{"title":"Move Prediction in Go with the Maximum Entropy Method","authors":"Nobuo Araki, Kazuhiro Yoshida, Yoshimasa Tsuruoka, Junichi Tsujii","doi":"10.1109/CIG.2007.368097","DOIUrl":"https://doi.org/10.1109/CIG.2007.368097","url":null,"abstract":"We address the problem of predicting moves in the board game of Go. We use the relative frequencies of local board patterns observed in game records to generate a ranked list of moves, and then apply the maximum entropy method (MEM) to the list to re-rank the moves. Move prediction is the task of selecting a small number of promising moves from all legal moves, and move prediction output can be used to improve the efficiency of the game tree search. The MEM enables us to make use of multiple overlapping features, while avoiding problems with data sparseness. Our system was trained on 20000 expert games and had 33.9% prediction accuracy in 500 expert games","PeriodicalId":365269,"journal":{"name":"2007 IEEE Symposium on Computational Intelligence and Games","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131307203","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}
T. Butter, Franz Rothlauf, Jörn Grahl, T. Hildenbrand, J. Arndt
Important advantages of genetic algorithms (GAs) are their ease of use, their wide applicability, and their good performance for a wide range of different problems. GAs are able to find good solutions for many problems even if the problem is complicated and its properties are not well known. In contrast, classical optimization approaches like linear programming or mixed integer linear programs (MILP) can only be applied to restricted types of problems as non-linearities of a problem that occur in many real-world applications can be modeled appropriately. This paper illustrates for an entertaining student game that GAs can easily be adapted to a problem where only limited knowledge about its properties and complexity are available and are able to solve the problem easily. Modeling the problem as a MILP and trying to solve it by using a standard MILP solver reveals that it is not solvable within reasonable time whereas GAs can solve it in a few seconds. The game studied is known to students as the so-called "beer-run". There are different teams that have to walk a certain distance and to carry a case of beer. When reaching the goal all beer must have been consumed by the group and the winner of the game is the fastest team. The goal of optimization algorithms is to determine a strategy that minimizes the time necessary to reach the goal. This problem was chosen as it is not well studied and allows to demonstrate the advantages of using metaheuristics like GAs in comparison to standard optimization methods like MILP solvers for problems of unknown structure and complexity
{"title":"Genetic Algorithms for Finding Optimal Strategies for a Student's Game","authors":"T. Butter, Franz Rothlauf, Jörn Grahl, T. Hildenbrand, J. Arndt","doi":"10.1109/CIG.2007.368119","DOIUrl":"https://doi.org/10.1109/CIG.2007.368119","url":null,"abstract":"Important advantages of genetic algorithms (GAs) are their ease of use, their wide applicability, and their good performance for a wide range of different problems. GAs are able to find good solutions for many problems even if the problem is complicated and its properties are not well known. In contrast, classical optimization approaches like linear programming or mixed integer linear programs (MILP) can only be applied to restricted types of problems as non-linearities of a problem that occur in many real-world applications can be modeled appropriately. This paper illustrates for an entertaining student game that GAs can easily be adapted to a problem where only limited knowledge about its properties and complexity are available and are able to solve the problem easily. Modeling the problem as a MILP and trying to solve it by using a standard MILP solver reveals that it is not solvable within reasonable time whereas GAs can solve it in a few seconds. The game studied is known to students as the so-called \"beer-run\". There are different teams that have to walk a certain distance and to carry a case of beer. When reaching the goal all beer must have been consumed by the group and the winner of the game is the fastest team. The goal of optimization algorithms is to determine a strategy that minimizes the time necessary to reach the goal. This problem was chosen as it is not well studied and allows to demonstrate the advantages of using metaheuristics like GAs in comparison to standard optimization methods like MILP solvers for problems of unknown structure and complexity","PeriodicalId":365269,"journal":{"name":"2007 IEEE Symposium on Computational Intelligence and Games","volume":"59 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116338546","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}
Marcelo Souza Pita, S. S. Madeiro, Fernando Buarque de Lima-Neto
Vidya is a strategy computer game, god-style that can be seen as a rich environment where virtual beings compete among themselves for natural resources and strive within the artificial ecosystem. Although in this game the player cannot directly control the intelligent agents, he can give some intuitions to them. Together with these intuitions the agents, called Jivas $the most developed species of the ecosystem, devise actions through evolutionary computation. The game allows also the observation of all interactions among the various beings inhabiting Vidya. Interactions happen in a quasi-autonomous manner which grants the game with an interesting dynamics. The evolved Jiva's intelligence, which build-up during the game, can be reused in other game scenarios. This work might help on further understanding of some emergent autonomous behaviors and parameterization of intelligent agents that live in closely coupled ecosystems.
{"title":"Vidya: A God Game Based on Intelligent Agents Whose Actions are Devised Through Evolutionary Computation","authors":"Marcelo Souza Pita, S. S. Madeiro, Fernando Buarque de Lima-Neto","doi":"10.1109/CIG.2007.368120","DOIUrl":"https://doi.org/10.1109/CIG.2007.368120","url":null,"abstract":"Vidya is a strategy computer game, god-style that can be seen as a rich environment where virtual beings compete among themselves for natural resources and strive within the artificial ecosystem. Although in this game the player cannot directly control the intelligent agents, he can give some intuitions to them. Together with these intuitions the agents, called Jivas $the most developed species of the ecosystem, devise actions through evolutionary computation. The game allows also the observation of all interactions among the various beings inhabiting Vidya. Interactions happen in a quasi-autonomous manner which grants the game with an interesting dynamics. The evolved Jiva's intelligence, which build-up during the game, can be reused in other game scenarios. This work might help on further understanding of some emergent autonomous behaviors and parameterization of intelligent agents that live in closely coupled ecosystems.","PeriodicalId":365269,"journal":{"name":"2007 IEEE Symposium on Computational Intelligence and Games","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2007-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114390016","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：