The Improvement For Crossover Algorithm Of GA

Genetic Algorithm (GA), is a kind of high efficiently searching arithmetic which according to the idea in the nature heredity of biology, and imitate evolve the way of the animate nature, and solution the problem space proceeds proceed together for high speed search, is a new overall situation optimize search for calculate way. It manhunt that abandoned the tradition manhunt in the method the point is little, certain etc. irregularity of direction, imitate nature life-span evolve the process, and adopt the artificial to evolve the way to solve the space proceeds with the randomization to target proceed together for the search. Because it is direct to proceeds to construction objects operation, the nonentity beg the differential coefficient and continue with function of limitative, fit to proceed inter-current handle, now already extensive application is in the artificial life, nerve network, ma- chine study, conveyance problem, calculator science, combination optimize, optimize scheduling, control adapt by self and semaphore transact etc. field, and also obtained biggest success at the actual applied.The work process of the Genetic Algorithm (GA) is: simulate that Darwinian heredity choice and imitate the evolve the process of living creature with the natural selection, first regard as the problem to an individual or chromosome in the community, and adopt the simple code technique to express the every kind of complicated construction, and code the each one to the sign string, and proceed operation community again and again base on genetics (heredity, cross with variation), according to the prearranged target adapt degree function to each individual proceeds evaluation, according to the fit to exist and superior winning bad inferior to wash out of evolve the rule, and guide with the direction of make ensure the manhunt, choice demand reasonable reflect" fit to exist" of the gene physical law, and the cross must to be as far as possible to inherit the next generation the advantage, adoption according to orientation the degree's choice strategy in the current group to choose the individual, and continuously get more fit of community, at the same time with the overall situation proceed together to search for the way to search for the optimize superior individual within community, such mimicry life of evolve the generation on behalf to evolve to go, until beg satisfy superior solution that request.This text heredity calculate way inside of cross the calculate way's improvement to proceed discussion and study.1. The principle and construction of the right first heredity calculate way proceeds study and researchIn GA, beg the problem's solution, to regard as superior solution for the community space look for the satisfying problem requesting or look like the manhunt process of superior solution. Orientation programming with adapt to the generous character and is two point contented to inherit calculate ways, orientation programming adoption imitate the choice mechanism of the nature biology with heredity calculate way if cross to wait with variation to operate to comes proceeds, using to guide the calculate way how to proceeds in the community spaces manhunt; Orientation the generous character then adopt the degree's worth method of calculation orientation to take the gauge of a candidate for election the good and bad that solve, in order to choice among them of superior solution proceed the next generation's heredity.The basic step that GA solves the problem can be described as follows:⑴First produced the candidate for election solution of an early community which has N candidates for election individual is called to early community;⑵Calculation community inside each candidate for election orientation degree that solve;⑶According to the candidate for election the orientation degree solving proceeds choice operation, if candidate for election that pick out solution satisfy the terminating of calculate way the term, then the calculate way terminate, otherwise continue⑷;⑷According to the cross probability, and solve in the individual in the community the candidate for election with the machine two to two for couples, proceed to cross operation with the born new candidate for election solution;⑸According to the variation probability, election of insides solution cluster each individual for variation operation with candidate from⑷;⑹Chooses for the new generation candidate for election the solution; Turn the⑵.While use the GA to beg to solve superior solution of the problem, make use of each individual within right and early community of coding technique to proceed to code first, then in the early community with the target function that the machine choose some code to proceed the combination, conduct and actions to evolve that point of departure, and calculation the each individual is worth, and is also individual orientation degree, and immediately after make use of to choose the mechanism to choose to adapt to the high individual( namely the excellent) of degree with the machine from the solution's community, and the conduct and actions breed the individual of the process front.( namely father age). Then, make use of to inherit the crossing of calculate way inside with two kinds of operations of variations, right choose the empress's individual (father age) to proceed to breed. Continuously repeat above evolve with breed the process, until satisfy the at an end term. Evolve the process till the last superior solution within generation, is to make use of to inherit the calculate way the solution the in- come of superior problem to of over the ending, from but real super winning and bad inferior to wash out in nature choice mechanism.It has choose calculate way, cross calculate way, variation calculate way etc. in the heredity calculate way, these calculate ways to hand over to substitute the usage, adopt which way to use is according to the concrete problem. In all calculate ways, choice that worth calculation of candidate for election community and orientation is very count for much, the candidate for election community come to a decision to leave superior solution far near, that which candidate for elections can worth result of orientation make sure then the community will be eliminated, which superior community will be left to be used as the new generation candidate for election the community to come to continuously multiply greatly the next generation, until get the best superior solution.2. Research crossover algorithm of Genetic AlgorithmCross the calculate way to include two basic contents: one is from be set upped in advance by the choice operate to become of form couples the mating pool, and form couples to individual with the machine, and press of cross the probability to decide which is right if the demand cross the operation. Two is set up the point of intersection combine at part of constructions that these points of intersection in front and back form couples the individual to the commutation.Now in common use and several kinds cross calculate way as follows:⑴The single point of crossover: while set up a point of intersection with the machine in the individual string, and crossing to the part that two individuals at before that point or empress proceeds to with each other change, born two new individuals.⑵Two points of crossover: In the individual string with part of proceedings that set two points of intersections, crosses the individuals in two points with each other change, then born two new individuals.⑶Many points crossover: Is based on expansion of two point's crossover.⑷The consistently crossover: According as the shield word to decide which part in two individuals to cross.⑸The even to cross: According as the stochastic figure that produce within machine of which the part to exchange.In these calculate ways, have no which calculate ways is an absolute better than another calculate way in any a kind of circumstance of, therefore at usage want to proceeds according to the actual circumstance the choice.3. Improvement crossed calculate wayThe textual improvement aims to crossover of two points crossover. The main mean is: conservancy oneself in two points switching part, avoiding the co- de repeating and losing, then process crossover operation, finally deletion outside the exchanging part. The realization of calculate way code adopt C computer language and use array and point to proceed.In the realizes process of the code, first consider to produce two group stochastic data, then certain two point of intersection for exchanging data, proceeds these two group data in point of intersections to the one self's replication oneself, save this substring after the gene's rear , thus can avoid at losing data because of commutation, and also can avoid the data's many ambulation for increase or delete proceed, then proceed commutation data in two point of intersections, finally delete outside points of intersection that repeat.Usage array is based on producing data easily, but deletion data is more bothering, and complex degree of calculate way is high a little; Usage point to realize checking seek is complex, need to point step to step to backwards proceeds to seek, but the calculate way lowered the complex degree, therefore at needle form that usage point the needle to realize that calculate way, first adopt array born with the machine data, then use point to proceed the increasing and deleting of data.The application of GA bring up in the each realm, this text inherit the calculate way to use at encryption and decryption, establish father on behalf as the original text, and treat posterity as the secret text, and putting it into practice is still have certain difficulty. But future at not far of belief, there is a larger development in this side with GA.In a word, in time of more than one year, pass by diligent theories and calculate way that study and research, crossover is already realized, from produce data to exchange data then to commutation data is already realized, and the fundamental realized the design's purpose.