基于优化的模糊决策算法在多AGV调度问题中的应用

为解决多AGV在复杂系统中的任务调度问题，以提高生产效率缩短任务周期为目标建立地面轨迹模型和AGV搬运耗时模型。采用优化的模糊决策算法通过最小化每台AGV的搬运耗时来最小化搬运作业的总耗时。在算法设计过程中，采用分布式控制思想，将AGV离目标点的距离，AGV的当前载重和AGV的当前电量作为因素集，将AGV执行此次搬运任务的合适度作为评价集，建立AGV模糊评价模型；通过该模型得出各台AGV对任务适合度的评价向量进而寻找一种映射将评价向量转化为值，调度问题被转化为最优化问题，其中采用切割补偿优化综合模型的因素权重选择，采用分模式链式比较减少AGV的计算量。通过多次仿真实验，可以验证本文算法在缩短任务的执行周期上具有一定的可行性与有效性，并且在解决复杂系统AGV状态多变的调度问题下展现了一定的鲁棒性和拓展性。

Application of Multiple Automated Guided Vehicles scheduling problems based on the optimal fuzzy decision algorithm

In order to solve the task scheduling problem of multiple AGV in a complex AGV scheduling system, The ground track model and the time model of AGV transporting material are established with the purpose of improving the production efficiency and shortening the period of task, By using the algorithm of optimal fuzzy decision to minimize the moving time of each AGV, the total time of the handling operation is also minimized. In process of algorithm design, this paper adopted the idea of distributed control. The distance between AGV and workstation as target, AGV''s current load and AGV''s current electricity will be as the factor set, the degree of appropriateness which expresses whether the certain AGV executes the current handling task will be as the evaluation set. Based on these two point, the AGV fuzzy evaluation model can be set up; Through this model, the evaluation vector of each AGV can be obtained, and then find a map in which the evaluation vector be converted to specific value, the scheduling problem can be regarded as a optimization problem. When choosing the weight vector in the fuzzy decision model, the method called cutting-compensation was applied to optimize the weight vector. When having the comparison of suitability between all the AGVs, The chain comparison base on differentiated pattern can reduce the amount of calculation of AGV. Through a lot of simulation experiments, it can be validated that the algorithm in shortening the period of task has certain feasibility and effectiveness, and the algorithm in solving complicated AGV scheduling problem where the state of AGV changes frequently shows certain robustness and expansibility.