Research On Multi Constraint Three Dimensional Packing Scheme Based On Genetic Algorithm

In recent years,the development of the logistics industry is in full swing.The most important index data in the logistics industry is the logistics cost,which includes all kinds of expenses of goods in the transportation process.It is the most concerned content for the logistics industry.Therefore,improving the efficiency of loading goods,reducing logistics costs,and increasing company revenue are the urgent problems to be solved in the current development of the logistics industry.In order to solve the above problems,this topic puts forward loading algorithms of the same size and different sizes,gives the multi-constraint cargo packing scheme,designs and implements a virtual visual loading system,gives the loading steps and loading instructions files.Firstly,combined with the current development status of the logistics industry,the packing algorithm theory and loading strategies at home and abroad are analyzed,the problems and deficiencies are summarized,the research content and key issues to be solved are put forward,and the classification of packing problems and the differences of related algorithms are summarized.Secondly,a three-dimensional packing problem is modeled with the maximal objective of space utilization,and a hierarchical loading heuristic algorithm is proposed for the samesized goods.The problem is converted into a two-dimensional packing problem after layering.The maximum number of placements in the plane is obtained by a five-block partitioning algorithm,and the optimal solution is obtained by linear programming between the layers.The loading algorithms of space selection,space partition and space merge are presented for different sizes of cargo loading.The above algorithm is tested with small,special and large cargo,and the space utilization rate is between 80% and 100%.At the same time,the actual loading sequence of goods is obtained by using the algorithm of generating loading sequence.Thirdly,the problem model is built to optimize the utilization of space and load for the constrained loading of goods,and the constraints of rotation and reversal of goods are also considered.For this problem model,genetic algorithm is used to solve the problem.A twosegment encoding including cargo number and placement status is designed.When the initial population is generated,the result obtained by heuristic loading strategy is coded into an initial individual,and then the initial individual is added to the initial population through piecewise random exchange replacement strategy.An adaptive crossover operator and an adaptive mutation operator are established by using a roulette algorithm for individual selection.The example data of Bray are used to validate that the whole cargo can be loaded in a single box,which takes 2 seconds to solve.The results show that the algorithm has obvious optimization effect for solving small,large and multi-constraint packing problems,and has certain validity for improving loading efficiency.Finally,based on the related algorithms,a virtual visual packing system is designed with three-tier architecture of ASP.NET in C# language.JQuery and three.js are used to achieve three-dimensional visual display of loading steps and processes.Loading guidance files are provided for users to carry out actual loading,and the feasibility and practicability of the system are illustrated by an example test.