Applying Coordination in Moving Target Search (MTS) Algorithm

Multi Agent System, where multiple agents concurrently and autonomously search to find a target, is regarded as a promising paradigm for future distributed computing. Agents often do not know the terrain in advance but automatically observe it within a certain range around it. In real time heuristic search such as Moving Target search (MTS), we have to commit a move within a limited search horizon or time and these heuristic search algorithms learn and improve their performance over successive problem trails. Multi Agent Moving Target Search (MAMTS) is a multi-agent version of Moving Target Search (MTS) algorithm where multiple agents concurrently and autonomously try to achieve a common goal. In this paper, we propose CMTS algorithm which uses color code for coordination among multiple agents. Each agent observes the color code of the state and selects the next move on the basis of this color; multiple agents, therefore, will find and travel distinct paths towards the goal state. In contrast, multiple agents may travel the same path in the original MAMTS. We have evaluated this coordination scheme on a large number of problems of varying difficulty generated randomly. Experimentally we have shown that as we increase the difficulty level CMTS out performs MTS. Key-Words: MTS, A*, CMTS, Online Search, Agents, Multi-agent, Coordination, Color Code, Real Time