The Assembly Sequence Planning Based On Improved Ant Colony Optimization

Assembly Sequence Planning (ASP) is the most important part of the Virtual Assembly Planning. It offers important basic information to Assembly Path Planning, Fixture Designing and so on. Essentially, Assembly Sequence Planning is a problem of NP ( Non-Deterministic Polynomial) combination planning. In this Paper, the Ant Colony Optimization is improved in order to solve the problems of Assembly Sequence Planning.The approach of assembly model is studied at first. In this work, assembly model is established using the approach of undirected graph, including the information of parts and the assembly connection information between parts. The assembly connection information is transformed to mathematical information, using assembly connection matrix and integrate interference matrix, on this basis, the method how to recognize the basic parts is introduced.Usually, there is some sub-assembly in an assembly, the correct recognition of sub-assembly can improve the parallelism of assembly operation. According to assembly connection information, the recognition algorithm of type I sub-assembly and type II sub-assembly is proposed in this work, and an example is introduced in order to inspect this algorithm.How to confirm an assembly sequence has a key impact on assembly sequence planning,in this paper, according to degree of assembly operation, redirection of assembly operation,tool change count of assembly operation, so a method of how to evaluate these assembly sequences is founded in this work.In this paper, there is a dissective analysis of basic Ant Colony Optimization. Basic Ant Colony Optimization is not fit for solving the problems of strong constrained condition, and it tends to appear local optimization. On this basis, a improved-Ant Colony Optimization is introduced, in which some aspects are improved: (1) The initial position of ant colony is the basic parts of an assembly, and in order to ensure the geometry feasibility of assembly sequences, the settings of ant colony's select range depends on assembly connection information. (2) In order to successfully avoid the local optimization, local pheromone is also updated while global pheromone is updated during the planning process.In the end of this paper, according to all the analysis above, the assembly sequence planning based on improved Ant Colony Optimization is completed. In order to verify the effectiveness of the assembly sequence planning system, a reducer is introduced as an example of assembly sequence planning.