| Time-cost-quality multi-objective optimization in construction project is a typicalNP-hard problem. Under certain constraints, its objective is to deal with the conflictbetween the objectives using some methods, obtain the optimal solution and determinethe optimal portfolio, to achieve the overall optimum of three major objectives. With therapid development of economy and engineering technology, there are increasinglynumber of large scale projects and complex technical systems. Therefore, constructionproject put forward higher requirements to the achievement of short duration, low-cost,and high-quality, which also makes the realization of the trade-off of time-cost-qualitymore important and difficult under the complicated engineering background.At present, during the solving methods of time-cost-quality multi-objectiveoptimization, mathematical methods have a large mount of computational effort andheuristic methods are lack of rigid enough, which are not applicable to solve thisproblem. The ability of meta-heuristic algorithms, in particular the evolutionaryalgorithms, to solve this type of problems has been widely demonstrated. Aimed at thetwo defects of current Evolutionary algorithms: dealing with candidate solutionsindifferently, not considering the potential of each individual solution, and withoutpaying much attention to the change in the surrounding environment, this paperintroduced a novel algorithm: Electimize. Based on the research and improvement ofbasic version, a new hybrid Electimize algorithm is proposed to solve multi-objectiveoptimization.Based on the elaborating of multi-objective optimization theory and according tothe preferences of decision makers and actual situation, this paper integrated theprimary objective method and linear weighted combination of methods, and establishedtime, cost single-objective optimization model and the time-cost-quality multi-goalsoptimization model under certain constraints; then, this paper expounded the basicprinciple Electimize algorithm, various elements and operation processes systematically,and coded the algorithm in Matlab. The optimization results of a classic case showedthe applicability and efficiency of the basic Electimize to solve multi-objectiveoptimization. In terms of ensuring global convergence and avoid local optima, thispaper made some appropriate improvements from the elitist strategy, geneticmanipulation, and division multiple swarm optimization strategy, and proposed a hybrid Electimize algorithm; In terms of the parameter settings, this paper obtained a parametercombinations to accelerate the convergence of the algorithm using a series ofoptimization experiments. Finally, use basic Electimize algorithm and hybrid Electimizealgorithms to solve a construction case, and then compare and analyze three results,including the result by particle swarm optimization. The results show that: the proposedhybrid Electimize algorithm can deal with construction project multi-objectiveoptimization with a higher level of application and convergence. |
