基于多变异分组遗传算法的多机协同作业静态任务分配

Static Task Allocation for Multi-machine Cooperation Based on Multi-variation Group Genetic Algorithm

摘要:

为解决农业机械（简称农机）多机协同作业前的任务分配问题，提出了基于多变异分组遗传算法的同种农机多机协同作业静态任务分配方法。首先，根据农机合作社实际作业模式，基于农机性能参数和任务参数，综合考虑机群的作业时间、作业油耗和路程代价等因素，建立了多机协同代价函数；根据多机协同作业特点，构建了多变异分组遗传算法，设计了两段式编码、分组交叉算子和多种变异算子，并建立了农机多机协同作业静态任务分配模型；通过仿真试验对比分析了算法的性能，并采用不同的代价权重进行了任务分配仿真试验；最后，采用不同的权重对实际深松作业进行任务分配试验，对本文提出算法进行了验证。结果表明：在不同权重下，基于多变异分组遗传算法的多机协同静态任务分配的机群代价比实际作业代价降低了29.48%~55.00%，选取合理权重的静态任务分配具有较高的分配效率和分配性能，能够满足实际多机协同作业中任务分配的要求。

Abstract:

Agricultural machinery cooperative operation often face the situation that multiple agricultural machines work together to complete multiple tasks in a working season. Reasonable task allocation before the operation can significantly improve the efficiency of the group’s operations and shorten the operating time.In order to solve the problem of static task allocation for multi-machine cooperation of the same kind of agricultural machinery,a static task allocation method of multi-machine cooperation based on multi-variation group genetic algorithm (MGGA)was studied. Firstly, according to the operation mode of agricultural machinery cooperatives and considering the operating time, the operating fuel consumption and the distance cost of the group comprehensive,a multi-machine cooperative cost function was established based on agricultural machine performance and task parameters. Secondly, based on the characteristics of multi-machine cooperative operation, a MGGA was designed and a static multi-machine cooperative task allocation model was established. Then, simulation experiment was carried out to compare and analyze the performance of the algorithm. And different weights were used for task allocation simulation experiments. Finally, practical experiment was carried out with different weights to verify the practicability of the algorithm by assigning tasks to actual deep loosening operations. Practical experiment results showed that the MGGA based task assignment multi-machine collaboration cost was 29.48%~55.00% lower than the actual work cost, which indicated that the MGGA based agricultural multi-machine cooperative static task allocation method had better distribution efficiency and higher distribution performance, and met the requirements of actual needs of agricultural multi-machine cooperative task allocation.