Solving the Heating Process Design Problem in Evolutionary Computation

摘要

In this study, we are concerned with a heating process design problem for a reactor filled with chemical material. The goal is to design a temperature profile such the reaction time is minimized and the mauufacturing constrains are satisfied. We propose a GA-based approach which can solve this design problem effectively. In order to accelerate our system performance, a distributed computing environment based on the master-slave model is suggested. The experimental results show that the system performance indeed can be speeded up near linearly for the number of nodes less than four in the network. However, for the computing network with more than three nodes, the communication cost grows up so quickly that a linear speed up can no be achieved.