Research Of Operation Control System Process Improvement Based On GA

Operational Control System is one of the subsystems in Securities Services system. As business grows, the efficiency of Operational Control system becomes more and more crucial. Operational Control system is made of hundreds of components. Separately optimizing on each component consumes so much effort that the development cycles and the cost would increase greatly. That would not be practical. General way is to have a preprocessing on all the modules. Firstly, we will focus on the system as a whole, and then some modules in the system would be merged. After that, the logic sequence would be considered. Then one important step for optimization is critical path searching. The critical path means the most time consuming chain in the whole system. The chain of critical path could give the system architect valuable information on optimizing the system.This dissertation mainly focuses on two problems. One is the critical path searching, another is simplified flow chart drawing. We propose GA (genetic algorithm) based algorithms for each problem.The mathematical model for the Operational Control System is a DAG (directed acyclic graph). GA is a random searching algorithm. Considering its large search space, we have some constraints before using GA. Our constraint is to partition all nodes to different layers according to their adjacency matrix. Besides that, we have a new object function. This object function is considered across and down. In across direction we try to keep the points are distributed as equally as possible while in down direction we try to retain the symmetry. The crossing number is also taken into account in the object function. And the encoding method is real number coded based on layered.The simulation result is given at the end based on MATLAB genetic arithmetic toolbox. Initial population's influence on the convergence time is analysis. Also the average value of object function is painted to show the convergence feature. For critical path searching, the convergence feature is also analyzed.