Task Allocation Strategy of Multi-Agent Based on ISODATA Algorithm

After an earthquake, it will cause great damage to urban road traffic, housing construction and people's life safety. It is the most important thing for assigning the rescue team to arrive at the disaster scene as soon as possible. But in an actual large-scale earthquake, these are quite a few difficulties for search and rescue task. For example, complex situation, long-distance communication blocking and the high risk for human rescue. In order to solve these complex and difficult problems. The paper proposes that ad hoc agents carry out rescue task, and assign agents to every area of the city by using the idea of clustering. Meanwhile, due to the shortage of traditional clustering methods like K-means, the paper proposes an allocation strategy of multi-agent based on ISODATA Algorithm, firstly cluster different urban environment adaptively, then assign search and rescue team to the corresponding region, the search and rescue team will carry out rescue at last. The experiments demonstrate that this method not only has a better performance compared with K-means, but also has a better performance in the whole rescue.