Parallel Computing Framework for Optimizing Construction Planning in Large-Scale Projects

Abstract

Available construction optimization models can be used to generate optimal tradeoffs between construction time and cost, however their application in optimizing large-scale projects is limited due to their extensive and impractical computational time requirements. This paper presents the development of a parallel computing framework in order to circumvent this limitation. The framework incorporates a multi-objective genetic algorithm module that identifies optimal trade-offs between construction time and cost; and a parallel computing module that distributes genetic algorithm computations over a network of processors. The performance of the framework is evaluated using 150 experiments that represent various combinations of project sizes and numbers of processors. The results of this analysis illustrate the robust capabilities of the developed parallel computing framework in terms of its efficiency in reducing the computational time requirements for large-scale construction optimization problems, and its effectiveness in obtaining high quality solutions identical to those generated by a single processor.