Optimization Of Muli-project Resource Based On Improved Self-adaptive Genetic Algorithm

Resource optimization is one of important parts of network programming, whichincludes resource leveling optimization and project scheduling problem under conditions oflimited resources. Since intelligent algorithms developed in recent years do not use anygradient information or any other auxiliary knowledge, and can be applied to thelarge-scaled and complicated problems, the genetic algorithms among intelligent algorithmshave become the main methods for resource optimization problems. A variety of geneticalgorithms for resource optimization have been proposed by many scholars.The existing genetic algorithms for resource optimization usually use some fixed controlparameters, which may resul in that the algorithm can only find a local optimum, thus isinefficient in searching the global optimum in the resource optimization process. Animproved self-adaptive genetic algorithm for muli-project resource optimization is presented in this paper. The adaptive strategy combines the individual life span based onthe fitness and the age to regulate population size and genetic operators, which can not onlyimprove the optimization efficiency but also prevent the "premature" effectively. In order toachieve the resource optimization objective under simulaneous muliple projects, we processall projects and mix their codes together and, in particular for muli-project scheduling withlimited resource, introduce the0-1matrix to represent the sequential relationship amongdifferent processes. As a resul, we can avoid drawing and merging the muli-project network.Finally, some instances taken from references are solved in MATLAB environment, theoptimization resuls show that our algorithm is effective on some resource optimizationproblems.