Study On The Decision Of Software Release Planning Based On Genetic Algorithm

Software release planning is known to be a cognitively and computationally difficult problem. Different kinds of uncertainties make it burdensome to formulate and solve the problem. The solution using evolution method is one method to solve these difficulties. After the formalization of the process being discussed, the ideas are provided to solve the problem. The basic principles include:determining the decision variables and the constraints; defining an explicit target of the release planning; obtaining a set of diversified and qualified solutions by the iterations of evolutional process; selecting the final solution from a set of qualified solutions by using the multi-attribute decision aid method. Each step should be finished through the interaction of the experts. Iterations of evolutional process will be actualized by genetic algorithm.Genetic algorithm is a kind of evolutionary algorithm, which is based on the biological theory of Darwinian evolution. It obeys the "the fittest will survive" principle, which is the core idea of the theory. The individual of the software release planning is treated as chromosome. We will use chromosome’s genetic variation and a series of simulated biological evolutions to seek the optimal solution.The main research content of this paper includes:1) To study domestic and foreign research status of the software release planning, and formalize the software release planning problems. The core of the formalization is a combination of software release planning problem in hard constraints and soft constraints. It is important for experts to take participation in this field while solving this problem.2) To probe into the theories of genetic algorithms, and describe important concept of genetic algorithms and operators which are used in the genetic algorithms. Based on these theories, we will take advantage of the genetic algorithms to actualize software release planning, construct a practical way to solve software release planning problem by them, and provide a viable operation to solve the problem.3) To achieve the core process of the release planning based on genetic algorithm by using Java language as a tool. An experiment will be finished with analog data. According to conditions of the experiment, some reasonable adjustment will be made to fit the needs of the planning.