Research On Optimization Of The Automobile Assembly Sequence Based On TSP

Automobile assembly is an important part of automobile production,which is mainly to assemble each automobile according to configuration,drive,power,color and other factors.Because of the different requirement of each automobile,for every bath of automobile,it will effectively reduce costs and increase production efficiency that searching an optimal assembly sequence.Based on the production of an enterprise's family automobile assembly line,we get the automobile assembly sequence problem to research.The main research contents include:(1)Bring out the problem and establish the mathematical model.According to the enterprise's production situation of automobile assembly line we bring out an assembly sequence problem that its goal is to minimize the times of switching automobile and the production requirements considered comprehensively include automobile configuration,drive,power and color.By Quantifying and defining the assembly requirements of automobile four types,and making statistics on arbitrary assembly sequences based on four types of automobile attribute,we establish the constraints and the multi-objective function to minimize the switching times.(2)Transform the basic mathematical model into the TSP model.The automobile to be assembled is regarded as the path vertex of the TSP problem,and the distance between any two vertices is the specified value of the constraints on adjacent automobiles by factors such as configuration,drive,power and color.Then the automobile assembly sequence problem is transformed into a multi-objective optimization TSP model.(3)Use genetic algorithm to solve the single-constraint TSP model.Firstly,the constrained factors such as configuration,drive,power and color are firstly considered separately,then the original model is simplified to several TSP model with only one kind of constraint.When using genetic algorithm to solve the problem,the real coding method is selected for the individual,the selecting strategy is tournament strategy,and the crossover and mutation strategy is combination of the reverse order,exchange and cyclic shift strategy,but the own constraint variables are brought in the fitness function.The population size of the algorithm is 800,the genetic algebra is 2000,the crossover probability is 0.8,the mutation probability is 0.1,and the total number of cars to beassembled is 461.By repeatedly solving in MATLAB,when only considering the drive constraints,the minimum switching times is 43,the number of consecutive four-wheel-drive automobiles is no more than 2,and the number of two-wheel-drive automobiles between two batches of four-wheel-drive automobiles is [10,50].When only considering the dynamic constraint,the minimum switching times is 30,the number of consecutive diesel automobiles is no more than 2,and the number of gasoline automobiles between two consecutive diesel automobiles is [10,72].When only considering the color constraint,the minimum switching times is 7,the number of continuous black automobiles is [67,70],and the number of automobiles of other colors between two consecutive batches of black automobiles is [29,81],while the mutual constraint of automobiles of other colors is also satisfied.All the results can completely satisfy the constraint conditions and reach the expected goal.(4)Use genetic algorithm to solve the complex constrained TSP model.By solving each TSP model with single constraint factor,the effective value range of each control parameter of the genetic algorithm is preliminarily verified.Considering all constraints comprehensively,the multi-objective TSP model with complex constraints is composed of 8 components.When using the genetic algorithm to solve the model,the value of each control parameter is the same as part(3),the minimum number of comprehensive switching is 143,46 times of driving switching,and the individual results include 24 times of power switching,7 times of color switching and 66 times of configuration switching.The results can satisfy the constraints and achieve the desired goal,and the algorithm runs stably.