Integrated Multiagent Course Search to Goal by Epsilon-Greedy Learning Strategy: Dual-Probability Approximation Searching

This paper presents a multistart, multiagent robust approach to search for a pattern (route to a goal), it uses dual probability approximation searching with the gradual decrease of a simple temperature parameter, which is initially warm and cools as the search converges. This is integrated with e-greedy learning, which is used to monitor changes in the optimal search and in the course. The resulting method has various new properties, including, for example, excellent goal convergence, wide initial searching, jumping obstacles, and active high curacy sensing. We present simulation results that validate this approach in the presence of obstacles.