Application Of Artificial Intelligence Algorithms In Component Retrieval

With the coming of the Internet Age, the software industry is facing a series of changes. Increasingly complex, shorter time requirements and the length of software code change from a few hundred lines a few decades ago to million lines now. These changes make the traditional software development methods can not meet the current demand of the software business. Software industry is taking more and more pressure. Component-based software development is an effective way to pass the software crisis, on the basis of some component model; it selects one or more software components from component library according to the demand, and then assembling these components together to form a system. It is a highly effective and quality procedure for constructing application software system. With the development of network technology, the number of component in component library is bigger and bigger. Generally, there are more than one component can meet user's demand, and if we let users make choice manually, it will undoubtedly increase the requirement for users, and also, it isn't conducive to the intelligent component assembly, and needless to say it will increase the cost of component based software development. So it is a key issue on how to select an optimal or approximate optimal component from the candidate quickly and effectively, so as to improve the success ratio of the component assembly.The thesis has three parts. The first part describes some basic concepts of component,the three popular component model COM,EJB and CORBA, and software quality metrics model. And then it introduces the difference and contact between component library and database system and their main functions. The second part introduces four component retrieval methods according to the different component description form. The third part proposes two component retrieval method using two artificial intelligence algorithms:genetic algorithm and tabu search algorithm. The beginning of this part analyzes the feasibility of two algorithms in component retrieval problem, and then it respectively describes the component retrieval method based on genetic algorithm and the tabu search algorithm. In the end, it compare and analyze the simulation results, the two methods are proven to be effective.The purpose and meaning of this thesis is that in the condition of the number of component in the component library is huge, and at the same time there are many components can meet user's demand, we can use artificial intelligence algorithms to find an optimal or approximate optimal component and feedback to the user. It can reduce the degree of user's participation, improve efficiency of component assembly and accelerate the development of intelligent component assembly.