It’s a researching hot spot for the route optimization and realtime dispatching when use AGV. In actual, people use:
AGV route and dispatching
1． Math planning method
Choosing the optimalizing task and route for AGV can be regarded as an issue of task dispatching, and the math planning method is the traditional method of first option to solve dispatching, the process is actually a best-way selecting under the resource being limited. The used methods includes integer optimization, dynamic programming, petri method, etc. These methods works well when in small scale dispatching, but by the increasing of dispatching scale, the time cost from problem solving increases like exponent, which limits the application for the large-scale realtime route optimization and dispatching of this method.
2． Simulation method
It’s the computing simulation for AGV dispatching solution by simulating the real dispatching environment.Users and researchers can use simulation to test, compare and monitor some dispatching solutions to change and choose dispatch strategy. The adopted methods are discrete event simulation, target facing simulation and three dimension simulation. There are many capable softwares for AGV simulating dispatching. The witness software can make a quick simulating module for 3D demonstration and result analyzing.
3． Artificial Intelligence Method
The AGV dispatching process can be described as the most optimized process that meets solution set searching when in limited situation for the artificial intelligence method. Its knowledge expressing technology contain human knowledge, also it use all searching technologies to provide a satisfying solution. The specific methods contain expert system method, genetic algorithm, heuristic algorithm, neural network algorithm. Expert system method are mostly used in real action, it transfers the experience of dispatching expert to the understandable and executable dispatching rules of the system. Also the collision counteracting technology is used for solving the rule expansion and collision of large-scale dispatching.
Due to the advantages of concurrent calculating, knowledge distribution storage and strong suitability, it has become a hopeful method to solve the large-scale AGV dispatching problems.Currently, TSP-NP problem has been solved by using neural network algorithm, when on solving, neural network can transfer the solving of combined optimization to energy function of discrete dynamics, to gain the solution of optimized problem by minimizing the energy function.
The genetic algorithm is a solution optimizing method to simulate the heredity and variation of process of biological evolution of nature.When solving the dispatching optimizing problem of AGV, the genetic algorithm will express numerous possible dispatching solution as proper chromosomes by coding, calculate the fitness (like the shortest route)of chromosomes, to hunt for the chromosomes with the greatest fitness by ongoing copying, crossing and variation, so it’s the optimized dispatching solution.
There might be some flaws by using only one way for the dispatching solution. Currently, it’s a research hot spot for putting many methods combined to solve AGV dispatching problems. For example, combine the expert system with genetic algorithm, put the expert knowledge into the forming of initial chromosome group to fasten the solving speed and improve quality.
