Abstract:

In order to analyze the vehicle configuration problem of intelligent workshop in customized equipment manufacturing enterprises, an optimal model with dual constraints which are system production capacity and order due-date is built to minimize the AGV investment costs. Because the optimal problem is a random nonlinear integer programming problem, and closed-form expressions of constraints and decision variables are not available, a tabu search algorithm based on queuing network is put forward. Firstly, based on the Markov theory, an improved decomposition of state space method is developed to analyze the queuing system with random batch processing and obtain the system performance measurers. Then an optimization algorithm is put forward to determine the optimal program of vehicle configuration. Finally, experiments to assess the effectiveness and accuracy of the proposed method are reported by comparing the results with simulation, and a case is studied to obtain the reasonable optimal program of vehicle configuration.

Key words: material handling system, Automatic Guided Vehicle（AGV）, queuing network, nonlinear programming

摘要：

针对定制型装备制造企业智能车间中物料搬运系统的AGV数量配置问题，以最小化AGV投资成本为目标，建立具有系统产能和订单交货期双重约束的优化模型。由于该优化问题是一个随机非线性的整数规划问题，且约束条件无法用决策变量的封闭形式表示，提出一种基于排队网求解性能指标值的禁忌搜索算法求解该问题。基于马尔可夫理论，提出拓展的状态空间分解法求解具有随机批量搬运特征的排队系统性能指标值。提出一种嵌入排队网模型的优化算法求解AGV数量配置的优化方案。通过算例求解与仿真对比验证所提方法的有效性和精确性，并分析案例确定优化的配置方案。

关键词: 物料储运系统, 自动导航小车（AGV）, 排队网络, 非线性规划