Study On Multi-Objective Optimization Of Risk Decision-making In Construction Project Based On Genetic Algorithm

Along with the expansion of the project scale, the upgrade of the technology and the dramatic change of social and economic environment, the risks increase obviously and their relationships become more and more complex. Risk management has become the crucial factor in the success of a project. However, risks may bring not only loss and disasters but also profits. The most important is how to manage the risks effectively. The result of decision-making which is the core of risk management will have a direct effect on the activity.In order to make the decision-making more scientific and logical, this paper commits to consider more objectives to achieve the optimization of risk decision-making of construction project based on risk identification and risk assessment, and apply the multi-objective genetic algorithm which is an effective method in solving multi-objective problems to the process of decision-making.Firstly this paper analyzes the status of current research on decision-making of construction project risk management and multi-objective evolutionary algorithm, and indicates the faultiness of the current research. Then it introduces correlative theories about risk decision-making, multi-objective optimization as well as principle and flow of genetic algorithm. This is prepared for the following text.This paper proposes a mathematical model for multi-objective decision-making of project risk management. This model is developed by a multi-objective genetic algorithm NSGA-Ⅱ. An application example is analyzed to illustrate the use of the model and demonstrate its capabilities in generation and visualizing optimization among risk score, risk management cost and risk loss. The result which is estimate by reliability proves that NSGA-Ⅱis good at obtain a converging and diverse Pareto front.A new selection operator and a new crossover operator are introduced into NSGA-Ⅱ. The improved method is also applied to the decision-making model which has been built. Analysis of the application example proves that the improved method is capable of improving convergence and variety of Pareto front when compared with ordinary NSGA-Ⅱ. Result of the case studies demonstrates not only the validity of the model and its method in solving multi-objective decision-making problems in the field of construction project risk management, but also their practical application value.