The Research Of Task Allocation Method Of Swarm Robots Based On The Dynamic Particle Bee Algorithm

The multi-robot system is a typical and artificial system composed of many robots that have small volume and simple structure.It is a trend that allocation is produced by swarm robots in the risk and narrow area that human and large equipment can not enter produced by the earthquake and mine.Reasonable task allocation is the precondition of it that swarm robots can complete the task quickly and efficiently.Aimed at different task objectives how fast distribute fitting robots and the number of allocated robots directly affects the efficiency of the multi-robot system of completing tasks.Inspired research thought by the phenomenon of bees,birds and other social creatures in the nature finding food fast through cooperation this paper put forward multi-robot task allocation method based on dynamic particle bees algorithm to improve the efficiency of task allocation in the process of searching in the unknown environment.Firstly,this paper analyzes the hierarchical structure of the heterogeneous multi-robot system and establishes the mathematical model of task allocation by combinatorial auction law.What's more,we improve the model as a mixed integer linear programming problem by introducing the sequence number of detected target to avoid big quantum circuit in the process of finding solution.Considering the performance value of every robot according to different task goal,the time of completing the task,priorities of each task and the distance of between robots and tasks and so on the fitness function of the optimization algorithm is established.Secondly,we optimize task allocation problem by the dynamic particle bee algorithm that combines the advantages of the birds and bees making the maximum of probability of robots selecting tasks as optimization objection.The method updates the speed and position of best bees and elite bees and scout bees by using the dynamic particle swarm algorithm to improve local search.So the proposed method solves the problem falling into local optimum easy of bee algorithm to make the optimizationeffect better.Then,at optimization stage of DPBA when dynamic particle swarm algorithm is in the local search we join neighborhood setting,which ensures update-location in range of maximum and minimum of the location in the process of local search to improve the efficiency of completing tasks.Finally,the paper designs the simulated experiment platform suitable for the problem of multi-robot task allocation and makes corresponding experiments at the simulated experiment platform.The experiments show the proposed method realizable.And the results also prove that the time of task allocation is shorter than that using bee algorithm and dynamic particle swarm algorithm and the error of the task assignment is smaller than that using bee algorithm and dynamic particle swarm algorithm and the convergence of task allocation is better than that using bee algorithm and dynamic particle swarm optimization.These verify the superiority of task allocation method of swarm robots based on the dynamic particle bee algorithm.