|
The methodology of the project-scheduling has been developed for several years after the CPM and PERT invented. All of them assumed that the resources are endless, so they cannot satisfy the limit of the resources. Due to the development of the technology and the more division of labors, the projects include more hundreds or thousands of jobs. Therefore, the scheduling of projects not only consider the priority of jobs, but also to think about the uncertain of the resources. It lead to the motivation of the research.
The research would take advantage of parallel multiple searches of genetic algorithm to provide the genetic algorithm-based scheduling system. In the solution process, enhancing the system’s efficiency, genetic algorithm will be utilized to achieve the global search and introduced the idea of using the immigration operator to strengthen the searching ability of algorithm.
Regarding to the setting of parameters in algorithm, different parameters in the research were designed form the experiment’s design to find out the most stable and suitable parameters set. Form the design of the fitness function, the research set up the reasonable weight for all the factors by the chooser in order to solve the multiple objectives that are mutually conflicted. In order to achieve the situation of reality, the fuzzy set theory will be utilized to represent the uncertainty in the amount of resources, using fuzzy mathematical programming to transform the model into the traditional linear programming model. The program and its results were tested by sample test in the PSPLIB. Then the program can find out the best solution. Finally, it is important to discuss the parameter of the algorithm; the relationship between the algorithm and the density of the project’s priority.
|