Research On The Problem Of Optimizing The Path Of Automated Warehouse

Automated warehouse, a kind of new technique, is comprehensiveengineering of sicience and technology about the perterage of matercalin the science of warehouse. With the development of modern technology,automated warehouse, which is an important part of Computer IntegratedManufacturing System (CIMS) and modern technique of logistics, willbecome more and more important.Although the application of automated warehouse has improved theefficiency greatly, there are still some difficult problems to be solvedin it. Among those, how to optimize the path problem is most critical,and also most important for the efficiency of automated warehouse.Thus, the paper focuses on optimizing the path problem of automatedwarehouse. As we know, the automated warehouses that have been put intouse are based on the fixed sheves. And we also know that the efficiencyof selecting fixed shelves. For the reasons mentioned above, the textis mainly about the selection of the optimum path for fixed shelves.In chapter 2, I have begun with the method of solving the pathproblem of automated warehouse by the means of Ant Colony Algorith (ACA).After that, I have deduced the target of this optimization (theselection of the shortest time) by analyzing the characteristics ofautomated warehouse in itself. Then, I have put forward the mathematicmodel and steps by which the problem can be solved. During that course,the imitation of real experiment has been made in a large scale. At theend of the chapter 2, I have compared this problem with the problem TSPto find out the differences between them.In chapter 3, I studied the problem by Genetic Algorithm (GA).Firstly, basic GA is put into use to abtain the mathematics modle andsteps by the description of the process of GA.Secondly, GA has been improved that of contents of this text unique.I put forward the concept of Relative Coefficient (RC), which indicateshow are two individuals alike. In this chapter, to assure the crossingpopulation, which is of great average relative coefficient (ARC) torecombine the cross population, but not to operate crossingly, I limitthe ARC. So, the efficience of the crossing operation is improvedgreatly and varieties of individuals come. It's certain that thelimited ARC is growing as the growing of the times of calculating toassure the calculation continues. I make a comparation at the end ofthis chapter.ACA and GA are combined to solve the path problem of automatedwarehouse in chapter 4. Compared by GA, the accuration of ACA is moreexcellent while the speed of it is more slowly, that is the basic pointof this combination.To begin with, I concluded the methods that the other researcherscombined the two algorithms by imitating of the two methods. The oneis the method of using ACA firstly and using GA secondly, while the otherone is using GA at first and then using ACA.After that, I put forward a new method of my own in which the twoalgorithms are combined by the means of parallel amend. The two waysof caculating are making at the same time, and they modify each othersome certain models to assure the acceleration of the optimization ofthe two algorithms, at a certain stage of calculation.In the end, I compared the all ways of the algorithms mentioned aboveand conclude:(1) The path problem of automated warehouse is the key factor ofworking efficiency of automated warehouse. What is more important isthat the problem of fixxed shelves needs careful reach. The applicationof ACA and GA, which prompted to solve the problem, needs continualimprovement.(2) The problem of optimizing the path of automated warehouse issimilar to TSP in some degree, but it has charaeristic of its own. Atfirst, the time of selecting is much shorter than that of the others.Secondly, the machine for stacking can move only on X and Y coordinates.Thirdly, there is more or less some distinction in speed. For thosereasons, the path in this problem is different form the problem of TSP.We can see from the picture that is for the goal of the shortest timeis similar to the path that is for the goal of the shortest distance.But they are different more or less.(3) ACA can solve the problem of optimize the path of automatedwarehouse. It's of the ability of finding out the best availableanswers and excellent efficience. It's complicated that a large numberof circulation are needed for seeking. Thus, It will take more time tofinish it. So I think more improvement should be made for ACA to beperfict.(4) The improved GA is not only of greater efficiency to the basicGA at the early stage of calculating, but also high efficiency can beassured in the final part of calculating. Besides, the time ofcalculation is no more than that of basic GA 15%. From these two points,we can say that the improvement for GA is successful.(5) There three methods (combination of GA and ACA) mentional inthe text have break through in a certain degree at the problem ofoptimizing the path of automated warehouse.(6) The method (combination of GA and ACA) discussed in the paperhas got much better efficiency than that of the two methods mentionedabove. For one thing, it has shortened the time of calculating. Foranother, the efficiency has greately improved. What is the mostimportant is that the advantages of ACA and GA are all strenghened inthe end.