Study on Multi-Agent Joint Parallel Computing for the Unknown Situation Path Explore / Exploit Search

• Main Authors: Chang-Kuo Chen, 陳章國
• Other Authors: Yee-MingChen
• Format: Others
• Language: zh-TW
• Online Access: http://ndltd.ncl.edu.tw/handle/84497063647445283787

碩士 === 元智大學 === 工業工程與管理學系 === 100 === Modern electronic warfare systems, the information loading and timeliness requirements is keep growing. How to quickly deal with these intelligence and the integration of computing under limited resources and time has become very important. MPI parallel computing provides decision-makers a way that dramatically reduce the computation time within the limited resources, MPI combine with digital pheromones particle swarm optimization by using the particle swarm to find out the optimal solutions through the swarms communication of the MPI parallel. The global optimal exploration then transits to a local exploit search. The local exploit search use Unity 3D and multi-agent with A * path finding method to search the shortest path and threat avoidance. We construct the agent behavior parameters for decision-makers adjust depth study and compare the simulation results under different scenarios.