Study on Probability Model of Multi-robot System on Task Allocation

The macroscopic probability modeling is a new way to describe and analyze the self-organizing multi-robot system. The mathematical method of stochastic process theory,matrix theory and linear algebra are used to model the task allocation of self-organizing multi-robot system and overcome the drawbacks of existing models,which has no scalability to the number of task types. The limit distribution of state-discrete time-continuous Markov chain is used as the theoretical result of task allocation,which takes advantage of predicting long-term behavior of robot system. The purpose of task allocation is to keep the proportion of the robots in arbitrary tasks to total robots being equal to that of performed tasks to total tasks. The simulation results show that the macroscopic probability model for multi-robot task allocation can complete the desired assignment of tasks.