The method based on q-learning path planning in migrating workflow

In a goal-oriented migrating workflow management system, each migrating instance is regarded as a mobile agent, the path planning for migrating instance is the path planning for mobile agent. The migrating workflow path is an ordered set of working positions that can achieve the sequence of goals carried by mobile agent. How to plan out a most efficient and most rational migrating path is one of the problems needs to be solved in the research of migrating workflow. This article puts forward a method that in a based-on social acquaintance network environment, mobile agent dynamically plan out a migrating work path by reinforcement learning. This method is suitable for the target-oriented migrating workflow management system, which can well solve the problem of the migrating path planning in the uncertain or partially observable environment of mobile agent.