Study On Multi-Objective Optimization Of Engineering Project Based On Genetic Algorithm

Nowadays, the research of network plan optimization can be mainly divided into three branches, time optimization, cost optimization and resource optimization. However, most literature only considers one or two of above objective indices which cannot fully satisfy the practical need of project management. On the other hand, project quality, which is another important objective of project management, has not been emphasized enough because it is difficult to be quantified. Although there are a few recent literature study on the optimization of time, cost and quality at one time, technically speaking, they are not multi-objective optimization at all because the method they use are to combine three objective function into one through linear weight factor, which has many disadvantages.The purpose of this paper is to build optimize models which can simultaneous consider all the above objective indices and to find effective methods to solve the model.The solution of multi-objective optimization problem are called Pareto optimal solutions(Pareto front) because there isn't any most optimized solution existed in such problems. NSGA-â…¡, which is an effective multi-objective genetic algorithm based on fast elitist non-dominated sorting, could search the feasible area in parallel mode by population evolution, and could get a lot of Pareto solutions without bias in one run.This paper first proposes a mathematical model of three-dimensional time-cost-quality optimization based on quantifying quality. This model is developed by NSGA-â…¡, a multi-objective genetic algorithm, which could avoid the subjective influence of weight factor. An application example is analyzed to illustrate the use of the model and demonstrate its capabilities in generation and visualizing optimal tradeoffs among construction time, cost and quality. The result proves that NSGA-â…¡is good at obtain a converging and diverse Pareto front.To consider another important objective factor, resource leveling, based on the aforesaid model, this paper added a function of resource leveling into it, and presents an advanced multi-objective optimization model that supports minimizing construction time, cost and resource, while maximizing its quality. Through embedded a genetic algorithm circle of resource leveling, the present model is implemented as a multi-objective genetic algorithm with a similarity-based mating restriction(SBMS) scheme to avoid the deterioration in the convergence speed to the Pareto front and to improve the diversity of non-dominated solutions. An application example is analyzed to illustrate the new method is capable of improving convergence and variety of Pareto front when compared with ordinary NSGA-â…¡.Cases studies with standard test problems are presented to demonstrate the performance of the models and methods could provide a new practical approach to multi-objective problems in construction project management.