Multi-variant Optimization Algorithm For Three Dimensional Container Loading Problem

The packing problem is very common in the field of industrial applications and in our daily life,in these such as industries cloth cutting,wood processing,logistics industry,storage and sharing of resources scheduling,scheduling etc.have very important application prospect.With the sharply developing of information technology.big data,data flow and data relations,the application of bin packing problem has been expanded into a new field.Three dimensional bin packing problem is a typical NP problem.For this problem,many scholars have made a great of efforts,but it is still difficult to get an effective algorithm.In recent years,some scholars put forward some algorithms to solve the three-dimensional packing problem.and achieved good solutions.However almost every algorithm has its own limitations,most of the algorithms are only applicable to fewer types of cases,while the problem have characteristics including that large number of cases,strong heterogeneous features,obviously the existing algorithms are facing enormous challenge.We need to rely on the characteristics of three-dimensional packing problem for practical application,in-depth study of the three-dimensional packing problem to find the appropriate algorithm.Multi-variant optimization algorithm is so make full use of the modern computer memories to memory some information about algorithm in the process of searching that increasing the sharing and the utilization efficiency of information.This dissertation investigates that three-dimensional container loading problem is solved by the multi-variant optimization algorithm.The multi-variant optimization algorithm applied random placement and partial adjustment gradually approximate the optimal solution.The random placement denotes that several cases of randomly selected are put into the container;the partial adjustment indicates that the sequence of the cases of random placement in the container are topically adjusted and optimized with objective function value.Then,the objective function value will gradually approximate the optimal value by recursively random replacement,partial adjustment and optimization,and we acquire a desirable three-dimensional container loading program.In order to verify the effectiveness and practicability of three-dimensional container problem with the multi-variant optimization algorithm,1000 groups of three-dimensional container loading problems that vary from BR1-BR10 are tested in this paper and acquire desirable results.Because of the easy implementation of the algorithm,the complexity of the algorithm is low,so the research of this paper is of great importance.The main work of this dissertation as follows:firstly,the basic principle of the multi-variant optimization algorithm is studied,analyzes the core idea and core elements of multivariate optimization algorithm,and the data structure of algorithm for storing and sharing information of optimization:then the basic process of multi-variant optimization algorithm and test method calculation example are given to describes the operation process of multi-variant optimization algorithm the more intuitive.Secondly,in order to verify the effectiveness and practicability of three-dimensional container problem with the multi-variant optimization algorithm,1000 groups of three-dimensional container loading problems that vary from BR1-BR10 are tested in this dissertation and acquire desirable results,and then the simulation of this algorithm compared with the existing algorithms for the packing efficiency.Thirdly,to facilitate the realization of the design and implementation of three-dimensional packing problem with multivariate optimization algorithm for the experimental platform for optimal performance evaluation,detailed examples from the experimental platform,operating instructions and experimental platform to test the three aspects to introduce the experimental platform.Fourthly,in the experimental platform,we do many tests with the test cases in BR1-BR10,and record the test results in real time.The conclusions of this dissertation are as follows:the algorithm based on the idea with random placement,partial adjustment,gradually approximate,proposed multi-variant optimization algorithm for solving the three-dimensional packing problem.Based on BR1-BR10 with total of 1000 test examples of simulation,analysis and comparison with other algorithms and solving the three-dimensional packing problem,proved the validity and feasibility of the algorithm to realize three-dimensional packing problem using the multivariate optimization algorithm;but there is still no parameter adaptive characteristics,the algorithm running time relative long.