基于改进蚁群算法的多AGV作业调度研究

发布时间：2018-06-01 07:13

本文选题：多AGV + 路径优化　；参考：《陕西科技大学》2017年硕士论文

【摘要】：多AGV作业是随着产业规模扩大、自动化程度提高、工序增多及柔性增大而逐渐得到越来越广泛的应用。多AGV作业调度问题对提高运输效率能够产生一定的影响,因此对它的研究在丰富多AGV作业调度的实际应用中具有重要的意义。本文重点研究AGV在车间装载零件过程的路径优化问题,多AGV在路口避碰的调度问题,建立以多AGV完成任务行驶路径最短为目标函数的数学模型,并提出改进蚁群算法对模型进行求解。本文主要研究内容如下:(1)在研究了国内外相关文献资料的基础上,针对基本蚁群算法迭代收敛缓慢、易跳出全局仅在局部展开最优解搜索等缺陷,提出一种应用局部和全局信息素同时更新的改进蚁群算法,并给出“最大最小交叉”策略,获得信息素新的更新方案,以改善传统算法的不足。(2)结合多AGV的工作流程,借助于交通流分配中交通阻抗的思想,考虑在路段和路口上产生的工作路径和辅助路径,分别从AGV作业调度的交通要求、运行速度的参数、行驶路径三方面研究。分析了实验场地中AGV行走的工作区域并将其转化为二维结构化空间,采用栅格标识直角坐标法建立了AGV行走空间的环境地图,为后期仿真实验做准备。(3)针对多AGV在路口避碰的调度问题,提出了高加低减调速解决路口冲突的策略,并建立了以多AGV完成任务行驶路径最短为目标函数的数学模型,应用上述改进蚁群算法进行求解,利用Matlab软件对蚁群算法和改进蚁群算法进行仿真实验,并对实验数据对比分析,结果表明改进蚁群算法在花费较小时间成本下可以不断溢出局部最优,能够在全局区域内寻找新的最优路径,且在最优解、平均解、最差解方面解的精度更高、性能更好。验证了改进蚁群算法在多AGV调度优化中的有效性,可为多AGV的实际调度作业提供参考。(4)为提高车间零件运输的自动化程度,实现最小投入获取最大效益的目标。本文以VB6.0为开发系统的平台,SQLserver2008为后台数据库,借助Matlab软件,设计开发多AGV调度管理系统,该系统不仅实现了零件装卸自动化管理的功能,同时利用混合编程将Matlab编写的改进蚁群优化算法在VB6.0中调用,给出零件装载的先后顺序,实现对多AGV调度管理系统的优化调度和管理,提高AGV在车间工作的效率。
[Abstract]:With the expansion of industrial scale, the degree of automation, the increase of working procedure and the increase of flexibility, multiple AGV operations have been more and more widely used. The multiple AGV job scheduling problem has a certain effect on improving the transportation efficiency, so the research on it is of great significance in the practical application of multi-AGV job scheduling. In this paper, we focus on the path optimization problem of AGV loading parts in workshop, the scheduling problem of multi-AGV at intersection, and establish a mathematical model which takes the shortest path of multi-AGV to complete the task as the objective function. An improved ant colony algorithm is proposed to solve the model. The main contents of this paper are as follows: (1) on the basis of studying the relevant literature at home and abroad, the basic ant colony algorithm has some shortcomings such as slow iterative convergence, easy to jump out of the global search for the local optimal solution, etc. An improved ant colony algorithm based on local and global pheromone updating is proposed, and a "maximum and minimum crossover" strategy is given to obtain a new updating scheme of pheromone to improve the traditional algorithm. With the aid of the idea of traffic impedance in traffic flow assignment, the work path and auxiliary path generated at road sections and intersections are considered. The traffic requirements of AGV job scheduling, the parameters of running speed, and the driving path are studied respectively. The working area of AGV walking in experimental site is analyzed and transformed into two dimensional structured space. The environmental map of AGV walking space is established by using grid marking rectangular coordinate method. In order to solve the problem of collision avoidance with multiple AGV, the paper puts forward a strategy to solve the conflict of intersection with high speed increase and low speed reduction, and establishes a mathematical model which takes the shortest path to complete the task with multiple AGV as the objective function. The above improved ant colony algorithm is used to solve the problem, and the Matlab software is used to simulate the ant colony algorithm and the improved ant colony algorithm, and the experimental data are compared and analyzed. The results show that the improved ant colony algorithm can continuously overflow the local optimum and find a new optimal path in the global region, and the accuracy and performance of the improved ant colony algorithm are higher in terms of the optimal solution, the average solution and the worst solution. The effectiveness of the improved ant colony algorithm in the optimization of multiple AGV scheduling is verified. It can provide a reference for the practical scheduling of multiple AGV) to improve the automation degree of the transportation of workshop parts and to achieve the goal of obtaining the maximum benefit from the minimum input. In this paper, we use VB6.0 as the development platform and SQL Server 2008 as the backstage database, with the help of Matlab software, we design and develop the multi-AGV scheduling management system. This system not only realizes the function of automatic part handling management. At the same time, the improved ant colony optimization algorithm written by Matlab is called in VB6.0 by hybrid programming, and the sequence of parts loading is given, which realizes the optimal scheduling and management of multi-AGV scheduling management system, and improves the efficiency of AGV work in the shop floor.
【学位授予单位】：陕西科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP23;TB497;TP18

【参考文献】