Study On Resource-constrained New Product Development Project Scheduling Problem

In order to keep the advanced advantage in the fierce market competition, enterprises continually seek changes to meet the new need. New product development has extremely vital significance to enterprises. Whether the product newly developed is successful not only depends on research and development capacity, but also on the level of project management, especially scheduling, for the market opportunity is fleeting, and the key of occupying the market is to take the lead to put the product into market. However, project scheduling can be influenced by resources. Thus, resources constrained project scheduling problem (RCPSP) becomes the research focus. One of the features of new product development project is duration's uncertainties, but existing research of RCPSP is based on the hypothesis of fixed durations and non-splittable activities, and pays little attention to research on RCPSP of uncertain durations and splittable activities is seldom. Therefore, this paper studies RCPSP on the basis of features of uncertain durations and splittable activities, so as to apply it into practice.Firstly, on the basis of reviewing project scheduling theory and its mathematical models, defines the features of development project, and establishes the mathematical model for RCPSP based on the features of uncertain durations and splittable activities.Secondly, discusses in detail the basic principles of particle swarm optimization (PSO)algorithm, as well as improved research achievements. Then, proposes a new improvement strategy on the basis of standard PSO model, and analyzes its application in resolving RCPSP of new product development project.Finally, applies the research achievement into practice of J Company's XX product development, and the result indicates that the model and improvement strategy are reasonable and effective.The research not only expands the field of RCPSP, but also designs a new improvement strategy for parameters on the basis of standard PSO model. This method can speed up the convergence rate of the algorithm, and avoid local optimum. It can also provide certain guidance for arranging the schedules of new product development project in a dynamic environment.