| With the development of industrial automation and intelligence,Automatic Guided Vehicle(AGV),as an important logistics equipment,is widely used in modern industrial production,especially in the field of material handling in the workshop,with broad application prospects.The workshop AGV scheduling system,as a key system to ensure the smooth progress of production processes,has been receiving attention from engineering and academia.There are some problems in the practical application of shop floor AGV scheduling system,such as low scheduling efficiency,unreasonable path planning,and unequal task allocation.Aiming at the above problems,this paper studies the workshop AGV scheduling system,aiming at improving the production efficiency of the production workshop,reducing operating costs,and improving enterprise profits.Specifically,the main research contents are as follows:(1)This paper studies the path planning problem of a single AGV,analyzes common map modeling methods,and selects the grid method for map modeling.Select Q-learning algorithm as the path planning algorithm for a single AGV,and propose an improved self-tuning Qlearning algorithm to solve the problems of slow convergence speed and unbalanced exploration.By improving the dynamic search factor,the relationship between utilization and exploration in the learning process is balanced,and heuristic ideas are used to initialize the Q value,effectively reducing ineffective iterations in the early stage and accelerating the convergence speed.At the same time,a new self-tuning estimator is introduced to correct the maximum deviation.Finally,a simulation experiment is designed using matlab software to verify the effectiveness of the improved idea.The improved algorithm is compared with other improved Q-learning path planning algorithms,demonstrating the advantages of the improved algorithm.(2)Study the scheduling strategy and principles of AGV scheduling system,select the minimum transportation distance as the scheduling principle for mathematical modeling,and use genetic algorithms to solve the task allocation problem.In order to solve the problem of slow convergence speed and easy to fall into local optima,the crossover factor and mutation factor have been improved to increase the diversity of the algorithm and effectively improve the search efficiency of the algorithm.Aiming at the problem that the continuous increase in cross factors and variation factors may lead to the elimination of excellent individuals,an optimal preservation strategy is adopted to effectively preserve the historical optimal individuals and obtain better optimization results.Through simulation experiments designed for different application scenarios,the effectiveness and advantages of the improved algorithm are verified from two aspects: path length and iteration efficiency.Aiming at the path conflict problem in multi AGV path planning,a method combining time window translation and re planning strategy is adopted to coordinate the conflict,which improves the operational efficiency of the scheduling system.(3)In order to solve the problem of low efficiency of the workshop AGV scheduling system,a multi AGV scheduling system based on the openTCS platform was built by analyzing the components and operating environment of the openTCS platform.The system can achieve vehicle scheduling,task allocation,and path planning functions.At the same time,it can monitor and control AGV vehicles,and visually display them on the page.In order to verify the feasibility of applying the algorithm studied in this paper to multiple AGV path planning,a workshop AGV scheduling experiment is designed to verify it.Through scheduling experiments,the effectiveness and superiority of the research method in this paper in multiple AGV planning,scheduling,and conflict handling are proved. |
