文档介绍:UTLogo100Stochastic ProgrammingSchool of Management, ic AlgorithmsKai YaoSchool of Management, UCASKai ******@ ProgrammingSchool of Management, ic AlgorithmsGA is a stochastic search method for optimization problems based on themechanics of natural selection and natural ics, ., survival of the?ttest. (Holland 1975)Representation StructureInitialization ProcessEvaluation FunctionSelection ProcessCrossover OperationMutation OperationGeneral ProcedureKai ******@ ProgrammingSchool of Management, UCASRepresentation StructureA key problem is how to encode a solutionx= (x1,x2,· · ·,xn) into a chro-mosomeV= (v1,v2,· · ·,vm). That is, we must construct a link betweena solution space and a coding ******@ ProgrammingSchool of Management, UCASAn ExampleLet (x1,x2,x3) be a solution vector in the solution space(x1+x22+x33= 1x1≥0,x2≥0,x3≥ encode the solution by a chromosome (v1,v2,v3) in the coding spacev1≥0,v2≥0,v3≥ the encoding and decoding processes are determined by the linkx1=v1v1+v2+v3,x2=rv2v1+v2+v3,x3=3rv3v1+v2+ ******@ ProgrammingSchool of Management, UCASInitialization Processpopsize: the number of chromosomes, for example,popsize= that the decision-maker can predetermine a region which containsthe optimal solution (not necessarily the whole feasible set).v1=Rand(l1,u1)v2=Rand(l2,u2)IfV= (v1,v2) is feasible,then accepted;otherwise ******@ ProgrammingSchool of Management, UCASEvaluation FunctionEval(V) is to assign a probability of reproduction to each ,V2,· · ·,Vpopsizebe the that the chromosomes can be rearranged from good to bad(., the better the chromosome is, the smaller the ordinal number it has).Rank-Based Evaluation FunctionEval(Vi) =a(1?a)i?1,i= 1,2,· · ·,popsizewhereais a given number, for example,a= ******@ucas.