Study On The Model And Solution Of Large-scale Assembly Line Balancing Problem

In the manufacture process of products,the assembly line balancing(ALB)is the core factor that restricts the production efficiency.How to solve this problem scientifically and efficiently is a continuous research hot topic for academic and engineering.This thesis takes the simple assembly line balancing problem(SALBP)as the research object,and conducts model and solving research on the minimization of the number of stations(SALBP-I)and the minimization of the production cycle problem(SALBP-II).Covering all scales of the problem involved in mainstream research at home and abroad,it provides model and algorithm support for establishing a general assembly line balancing problem solving system.First of all,the establishment of mathematical models for two types of assembly line balancing problems(SALBP-I and SALBP-II)is studied and mixed integer programming(MIP)models for two types of problems are established.After the model is established,the computer expression of the model is performed by the LeapMS modeling language,and the model is packaged by the C++ language.Finally,the solution is solved by the Cplex solver,and the solution is generalized,modularized,and automated.The SALBP-Data-Sets,an international standard benchmark set of the questions,is used to test the operation.The test results show that the exact solution of about 70%of the standard instances can be obtained by the method.Compared with the recent literature,the method of this thesis is more efficient,the structure and implementation are simpler and more intuitive.Secondly,the integer programming method for the problem is difficult to solve large-scale complex problems,a multi-population genetic algorithm is used to solve the assembly line balancing problem.The traditional genetic algorithm is improved to solve the problem.It includes the following:proposed the initial population generation based on random topological sorting,proposed a different decoding algorithm for the two types of problems,and designed a variety of genetic algorithm operations such as selection,crossover,mutation and excellent individual population migration algorithm strategies.The instances test show that there is a certain advantages over the recent literature when solving the nearly optimal solution of the large-scale assembly line balancing problem for the proposed algorithm.Finally,an assembly line balancing system based on an effective algorithm is established,and a graphical user interface of the system is designed to visualize the optimization results of the assembly line balancing problem.The two solution methods are integrated in the system to achieve multi-algorithm solution to the balance of the assembly line.The method proposed in this thesis can be used to solve small,medium-scale problems and large-scale problems at the same time.The solution scale has a wide applicability.The algorithm test covers all scales of instances in the international standard benchmark.The stability of the algorithm is good,and better feasible solutions can be found for the standard benchmark set.The research results of this thesis have a good reference value for the establishment of a general solution system for the assembly line balancing problem.