Study On Multi-objective Optimization Of Construction Project Based On Immune Genetic PSO

For the current development of information technology and its wide application inproject management, the quantification process for each objective in the project hasbecome particularly important. In addition, people have started to pay more attentionon how to optimize the benefit when taking multiple objectives into consideration,which include quality, schedule, cost, safety, environmental protection and so on.Therefore, the introduction of intelligent algorithm to achieve multi-objectiveintegrated optimization has become a hot research topic in both theoretical and appliedfields.The traditional optimization model takes only three factors into consideration,which are schedule, quality and cost. Considering the actual situation, safety andenvironmental factor, were added into the traditional model. Therefore, a novelfive-objective optimization model, i.e schedule, quality, cost, safety and environmentalfactor, was established in this paper. In this new model, based on the existingresearches about quantification of the project quality and actual “student syndrome”phenomenon in the construction project implementation, the index of standard for eachprocess was quantified. In addition, according to the reliability theory of processnetwork system, the measurement for project quality standard was set up, whichovercame the disadvantage of determining weight coefficients in a subjective manner.The Cost-Duration model was proposed for this project. In this model, the cost of thewhole project was divided into direct costs, indirect costs and project tardinesscompensation costs. Since safety is benefit, the marginal benefit theory was introducedinto safety input-output model. In this model, based on the theory of marginal benefitprojects. Moreover, the green degree model to evaluate environment impacts was alsoestablished based on LCA theory and life cycle of residential green construction. This paper analyzed the basic principles of immune genetic algorithm and particleswarm algorithm and also compared their advantages and disadvantages. According tothe PSO algorithm advantages of "self-improving" ability and learning from "others",the crossover and mutation mechanism of immune genetic algorithm were introducedinto the particle swarm algorithm. In addition, combination with the mechanism ofimmune memory maintained the population diverstity. This is our method to createimmune genetic particle swarm optimization algorithm. This algorithm can avoidpremature convergence of the algorithm, escape from local optima and then quicklyconverge to the global optimal solution, in order to establish the foundation to get thesolution for multi-objective optimization.In the last section of this thesis, this model was applied in a case study. Thesolution for multi-objecitve optimization was obtained via the IGPSO modelprogrammed in Matlab. The result shows, its feasibility and effectiveness inengineering problems.