The Implement Of Genetic Quantum Algorithm On Geometric Constraint Solving

This technology of CAD (Computer Aided Design) is a new kind of developing comprehensive computer application system technology which developed rapidly over the past 40 years. In recent years, CAD technology is applied to each field of industrial production extensively. As important technology of CAD, parameterization design becomes an effective method of the initial design, products modeling and revise, many scheme compare design dynamically during the process of the production design relying on powerful sketch map design, dimension drive and revise such functions as the figure, etc. And now, people pay more and more attention to it. Whether the parameterization design capacity is strong or not, it already becomes one of the important signs of weighing modern CAD systems. Geometric constraint solving technology is on the basis of constraint parameter design central technology. Whether the technology good and ripe or not will be the key of weighing one design system fine or not on the basis of parameter constraint system. Around the important problem of how to establish the method of parameterization design and geometry constraint model, domestic and international experts have carried on a large amount of research work, and have got a lot of effective achievements with practice and theory.The present solving methods can be summarized into: whole solving approach, sparse matrix approach, connect analytic approach, stipulation construction approach, constraint propagation approach, symbol algebra approach and auxiliary line approach etc.Viewd of the different angles, these methods improves the constraint solving capability of the CAD greatly. However, because of their limit, there are a lot of problems will be solved. For instance, numerical iteration method can solve one problem in the answer space, but for the multi-answer in the answer space, it can't compare. Because of the sensitive for the initial value, improper initial value will lead to user unexpected answer converged by the arithmetic. Still, there is the problem of rapidness and whole problem of the arithmetic convergence and the problems of under-constraint and over-constraint in the geometry constraints, etc.This thesis attempts to do breakthrough and innovation discussed in above-mentioned problems, so introduced a kind of new method of geometry constraint to improve the ability based on geometry constraint modeling, and expands the application level of existing CAD system. This new method combines the thought of quantum calculating with genetic algorithm to produce the genetic quantum algorithm.The genetic algorithm is a part of evolution algorithm, which now has been used widely in project optimize, signal disposal, pattern-recognition, administrative decision,intelligence system design and artificial life field,etc at present. But genetic algorithm has obvious shortcoming that is convergence speed slow and immaturity convergence.Because in the traditional genetic algorithm, although the operator of evolution can reflect searching for and information exchange between the individuals in the colony; it is the one that doesn't have evolution of thinking without the effective using the history of evolution.Fact proved that introduced good guide mechanism can reflect the rule of evolution, which can be used to guide the following evolution, make the algorithm have certain intelligent and improve the efficiency of the algorithm .Genetic quantum algorithm adopts quantum bit signifying chromosome, quantum chromosome can signify the state of superposition, make this algorithm have better colony variety than the genetic algorithm. Evolution variation of the quantum chromosome can introduct instruction message conveniently and observe the abundant colony at random. Furthermore, quantum crosses can overcome algorithm the early-maturing effectively. With α, βtending toward 0 or 1, the quantum chromosome disappears in a state, and variety disappears at this moment, the algorithm becomes convergent. The genetic quantum algorithm compares with traditional genetic algorithm on solving optimizing problems. There is great improvement both in convergence of the speed and seeking excellent ability. This paper realizes genetic quantum algorithm, and compares with the practice running result genetic algorithm through evaluating the concrete function extremum, validates the fact that genetic quantum algorithm is really better than the traditional algorithm in solving the evaluation optimization problem, and its convergence speed is improved a little. On solving geometry constraint problem, we convert the constraint problem into the equation group in this paper. We will get the optimization model by the simple summation of the equation absolute value, while other most optimization algorithms utilize the square summation of the constraint equation group from getting the model. Compared with these constraint models, the constraint model in this paper is simpler and can make the calculation amount reduce greatly. Then use...