Solutions SHaring in Multi-agent System
Hanbing Deng, Xia Zhang, Jiren Liu, Qian Zhang · 2012
In this paper, we will do some researches on the relationship between action sequence and system efficiency in multi-agent system.Usually, when an agent gets a task, it will calculate all the feasible action sequences and choose the best one whether the task has used to appeared in this system before.It wastes a lot of time on re-calculating and recomparing.So we propose solution sharing mechanism to make all the agents share the same solution space which keeps the current optimal plan for tasks.After achieved new tasks, each agent will have only one calculation and comparison, and then choose the better current action sequence.The reduction of time and better action sequence will improve the efficiency of multi-agent system.