A Study on the Application of Genetic Algorithm to Batch Processes Scheduling System for Power load forecasting Optimization

• Main Authors: William Wei-Yuan, 劉偉遠
• Other Authors: none
• Format: Others
• Language: zh-TW
• Published: 2003
• Online Access: http://ndltd.ncl.edu.tw/handle/73674466536148884641

碩士 === 國立高雄第一科技大學 === 機械與自動化工程所 === 92 === Due to the global competition, enterprises are not only reduce manpower but also improve process or conserve energy efficiently. The electricity-demand-control system is a powerful means among load management. In this research, the Genetic Algorithm (GA) is used to forecast electricity consumption of chemical batch processes production. By using the production schedule to balance the load distribution, it not only depresses the peak load and advance off-peak load, but also furthermore avoids power load from fluctuating. It provides a tool to monitor and manage electricity usage reasonably. In the past decades, the research of production scheduling mostly focused on the solution of minimizing makespan, minimizing mean flow-time, minimizing mean tardiness and minimizing maximum tardiness, and so on. But this thesis is mainly focused on the application of Genetic Algorithm (GA) on chemical batch processes production scheduling, in which the GA core technology of scheduling is implemented by way of modularization in order to make batch production scheduling effective and optimized. Taguchi method of the experimental design is employed to optimize system parameters in the application of electricity forecasting by using the Genetic Algorithm. Finally, scheduled production orders are verified against expert system and prove that power load curve is smoother than the resulting schedule generated by the expert system. The results show that parameters set 900 of population size, 0.8 of crossover rate, 0.3 of mutation rate and 800 of evolve generation is optimal for this study.