막힌 부분을 갖는 네트워크 내 효과적인 경로 탐색을 위한 유전 알고리즘 적용
Applying Genetic Algorithm for Exploring the Effective Path in Network with Dead Ends
0
조회
0
다운로드
1
인용
0
참고문헌
키워드
초록
자연계의 진화 과정을 모방하는 유전 알고리즘은 다양한 조합 최적화와 같은 NP-hard 문제의 해를 탐색하는데 매우 유용한 도구이다. 본 논문은 네트워크 내에 존재하는 두 노드 사이의 최단 경로를 구하는 문제 풀이를 위하여 유전 알고리즘을 적용하고자 하며, 특히 실행가능한 경로가 희소한 네트워크를 다룬다. 제안하는 유전 알고리즘은 적절히 선별된 노드의 다음 방문 노드를 명시하여 해를 표현하며, 별도의 복구 작업을 포함하지 않는다. 이러한 표현 방법을 사용할 경우, 해 집단 내에 실행불가능한 경로나 순환이 있는 경로가 포함될 수 있으나, 이러한 경로들은 적절히 설계된 적합도 함수에 의해 처리된다. 결과적으로, 제안하는 유전 알고리즘은 순환이나 막다른 노드가 많은 네트워크에서 효과적이고 실행가능한 경로를 탐색하는데 특히 유용할 것으로 생각된다. 잘 알려진 강 건너기 문제 중 하나인 선교사와 식인종 문제를 통해 실험한 결과, 제안하는 유전 알고리즘을 통해 전체 구조를 분석하기 어려운 네트워크에 속한 두 노드 간의 경로를 효과적으로 구할 수 있음을 알 수 있었다.
Abstract
The genetic algorithms that mimic the mechanism of the natural evolution is very useful tool for exploring the solutionsof the NP-hard problems such as a variety of combinatorial optimization problems. This paper aims to apply the genetic algorithm to solve the shortest path problem, and especially, the network with few feasible paths is considered. The proposed algorithm identifies the relevant nodes at first, and represents a solution by specifying their adjacent nodes to visit next. Moreover, any additional repairing procedure is considered. This simple encoding can make the population to include infeasible paths or paths with cycles, however, such paths can be deal with by the properly designed fitness function. Consequently, it is expected that the proposed genetic algorithm will be useful to explore the effective feasible paths in the networks with cycles or dead-end nodes, especially. The result of the experiment with the missionaries and cannibals problem, a well-known river-crossing problem, reveals that a path between two nodes in a network of which entire structure is hard to be analyzed can be obtained effectively by the proposed genetic algorithm.
참고문헌 (12)
- Gen, M., "Genetic Algorithms for Solving Shortest Path Problems", Proceedings of Congress on Evolutionary Computation 2004, pp.835~838 (2004)
- Ahn, C.W, "A Genetic Algorithm for Shortest Path Routing Problem and the Sizing of Populations", IEEE Transactions on Evolutionary Computation, 6(6), pp.566~579
- Hu, Y, "A Knowledge Based Genetic Algorithm for Path Planning in Unstructured Mobile Robot Environments", Proceedings of the 2004 IEEE International Conference on Robotics and Biomimetrics, pp.767~772
- Su, S., "Learning of a Maze Using a Genetic Algorithm", Proceedings of the International Conference on Industrial Electronics, Control, and Instrumentation, pp.376~379 (1993)
- Holland, J. H., "Genetic Algorithm", Scientific American, 266, pp.44~50 (1992)
- Elshamli, A., "Genetic Algorithm for Dynamic Path Planning", Proceedings of the Canadian Conference on Electrical and Computer Engineering, pp.677~680 (2004)
- Li, S., "Efficient Path Planning Method based on Genetic Algorithm Combining Path Network", Proceedings of the 4th International Conference on Genetic and Evolutionary Computing, pp.194~197 (2010)
- Inagaki, J., "A Genetic Algorithm for Determining Multiple Routes and Its Applications", Proceedings of the IEEE International Symposium on Circuits and Systems, pp.137~140 (1999)
- Carrick, C, "An Evaluation of Genetic Algorithm Solutions in Optimization and Machine Learning", Proceedings of the 21st Annual Conference of Canadian Association for Information Sciences, pp.224~231
- Baba, N, "Genetic Algorithm Applied to Maze Passing Problem of Mobile Robot - A Comparison with the Learning Performance of the Hierarchical Structure Stochastic Automata", Proceedings of the 1994 IEEE International Conference on Neural Networks, pp.2690~2695
- Gordon, V.S., "Evolving Sparse Direction Maps for Maze Pathfinding", Proceedings of Congress on Evolutionary Computation 2004, pp.835~838