Man-day Based Network Engineering Proiect Scheduling Optimization Study Using Particle Swarm Optimization Scheme

• Main Authors: Houng-Chang Tsai, 蔡鴻章
• Other Authors: Dr. Ruey-Maw Chen
• Format: Others
• Language: zh-TW
• Published: 2010
• Online Access: http://ndltd.ncl.edu.tw/handle/57621368607932198410

碩士 === 國立勤益科技大學 === 電子工程系 === 98 === For the smooth implementation of the network constructions project and maximize profits, a well estimated project schedule (construction project scheduling) is important. Good or bad scheduling will seriously affect the cost paid and profits earned; therefore construction project scheduling becomes the important factor in project success. Project planning in early day, different construction components were commissioned to different professional engineering firms for the facilities planning and design drawings, and then estimates construction period. All professional engineering firms, based on their accumulated professional experience, use the Gantt chart to schedule all the details of the construction and fill in corresponding construction time and manpower to figure out the required construction schedules and the labor costs. Recently, almost all network construction cases are implemented according to project management model, and make use of Microsoft office project tools for project scheduling. However, in Taiwan, most companies are used to use “Man-day” as an estimation unit. Hence, project scheduling remains inseparable from the rule of thumb even Microsoft office project tools are applied to estimate the construction period and the construction labor costs. Restated, experience decides everything. To enhance the estimation mechanism of project scheduling and to give project manager having reference guides during the time of construction project scheduling, this study proposes an improved group intelligence beased particle swarm optimization (PSO) to perform the optimization of man-day network construction and then to calculate the actual needs of labor. This work searches the network construction project scheduling analogous to a well known multi-mode resource-constrained project scheduling problem (MRCPSP) since “Man-day” of scheduling involves various combinations (modes). Accordingly, in order to find out the optimal network construction schedule, this study has to decide the mode, then scheduling. Additionly, the improved particle swarm optimization used in this study consists of two particle swarms to solve the project scheduling, they are forward and backward particle swarms. Meanwhile, a Gbest ratio (GR) is designed for the particle velocity update, i.e., to use Gbest mode to update velocity probability rate is GR; then, to use lbest mode to update velocity probability rate is (1-GR). This study uses the real network constructuion cases for experiment and to compare the simulation results with the output of labor estimation system of Customer relationship management, CRM. Moreover, further analysis is also given. The experimental results confirm that the proposed design based on particle swarm can effectively solve multi-mode resource constrained project scheduling problem, and optimize scheduling for network construction cases.