The handling of a single AGV is very simple. How to arrange multiple AGVs reasonably to maximize their work efficiency requires the use of an AGV scheduling system. Once the AGV scheduling system specifies a route for the AGV, the on-board controller on the AGV will complete specific motion control tasks according to the instructions, such as how much speed is maintained, how the wheels turn when turning at an intersection, and so on. Therefore, in a complete large-scale system, the AGV dispatching system is located between the upper control system and the lower control system, and it functions as a housekeeper. For multiple AGVs, the scheduling problem becomes very difficult, and the greater the number of AGVs, the greater the difficulty. Therefore, the dispatch system has become a key technology in the AGV industry.
The number of AGVs that can be dispatched has also become an important indicator for judging whether a dispatching system is strong or not. In this regard, Miklimi said: At present, there are not many AGV companies with self-developed AGV scheduling systems, and there are only a handful of AGV companies that can truly optimize the scheduling system and use the AGV system with the highest efficiency. The software functions of the AGV dispatching system independently developed by Mikroma include: task management, vehicle status monitoring, route planning, map editing, database query, etc.
AGV route optimization and real-time scheduling are a hot research topic in the current AGV field. There are mainly three commonly used methods:
1. Mathematical Programming Method
AGV selects the best task and the best path, which can be summarized as a task scheduling problem. The methods in actual use mainly include integer programming, dynamic programming, petri method and so on. In the case of small-scale dispatch, this kind of method can get better results, but as the scale of dispatch increases, the time spent in solving the problem increases exponentially, which limits the application of this method in responsible, large-scale real-time route optimization and dispatch .
2. Simulation method
The simulation method simulates the implementation of a scheduling scheme of AGV by computer simulation by modeling the actual scheduling environment. The methods used in practice are discrete event simulation method, object-oriented simulation method and 3D simulation technology.
3. Artificial intelligence methods
The artificial intelligence method describes the AGV scheduling process as a process of searching for the optimal solution in the solution set that satisfies the constraints. It uses knowledge representation technology to include human knowledge, and at the same time uses various search techniques to try to give a satisfactory solution. Specific methods include expert system methods, genetic algorithms, heuristic algorithms, and neural network algorithms.
The scheduling of multiple AGVs needs to plan the paths of different AGVs, so we first understand the existing path planning methods. AGV is an obedient fool. If the scheduling system wants an AGV to walk from A to B, it cannot simply give the AGV the information of sites A and B, but tell the AGV the complete path between A and B. At present, “graph” (mathematical concept of graph) is commonly used to model the driving space of AGV. The “graph” is composed of nodes and edges. Therefore, the driving path of the AGV can be expressed as a series of adjacent nodes.