Multi-robot Task Allocation Based On Fitness And Ant Colony Scattered Search

With the complexity of social production activities and the reduction of the labor force of the population,the application of multi-robot systems occupies an increasingly important position in our production and life.Multi-robot task allocation is an important research direction in the field of multi-robot systems.It is a prerequisite for multi-robot systems to perform tasks,and is also the key to efficient system performance.It is the core of organizing simple robots to complete complex tasks.Multi-robot task allocation is a classic combinatorial optimization problem,which is also an NP hard problem.How to construct a suitable and efficient optimization algorithm according to the task characteristics is worthy of research and development.This dissertation focuses on the research of the allocation problem of large-scale hybrid tasks.The algorithm mainly adopts automatic task classification and task allocation structure to optimize the allocation.In hybrid multitasking,an improved fitness modeling method is used to automatically classify and preliminarily assign tasks.In this process,the robot automatically screens single robot tasks(loose tasks)and multi-robot tasks(tightly coupled tasks),and the use of improved fitness for loose task assignments,can not only achieve the optimization of loose task assignment,but also make full use of the robot's autonomous selectivity to reduce the communication consumption between robots and improve computational efficiency.For tightly coupled tasks,the upper layer uses ant colony algorithm to optimize the task allocation.The ant colony algorithm uses grouping optimization to reduce the overlap of robots between alliance solutions and reduce the waiting time for execution.The union solution of the lower tightly coupled task is used.Using the improved scatter search idea to solve the task alliance,that is,according to the fewer robots performing the task,the communication,collision avoidance,and cooperative control between the robots will be reduced.The robot alliance solution is obtained according to the idea of integer coding.And through the establishment of alliance reference set,to continuously optimize the solution set of the updated reference set,to ensure that the alliance solution can better search.