Stochastic time–cost tradeoff analysis: A distribution-free approach with focus on correlation and stochastic dominance

The present study explores how correlated Monte Carlo simulation (MCS) coupled with a multiobjective particle swarm optimization algorithm can expand the time–cost tradeoff analysis in the presence of uncertainty. The goal of the proposed framework is to find the optimal set of activity options, whose objectives are evaluated as value-at-risk measures of project duration and total cost. The proposed framework incorporates the Gaussian copula into MCS to treat statistical dependence between uncertain variables, with no restriction on the estimation process and distribution type. This paper elucidates the definition of stochastic dominance relations, based on which a decision rule is established to prescreen dominated solutions so as to alleviate computational burden. A practical project has been used to validate the proposed framework by comparisons with enumeration and NSGA-II (non-dominated sorting genetic algorithm). In addition to nondominated solutions, the proposed framework provides insightful risk assessments.

Research Highlights► We integrate correlated Monte Carlo simulation with multiobjective Particle Swarm Optimization. ► We model statistical dependence between uncertain variables. ► Stochastic dominance and a prescreening strategy are proposed to increase efficiency. ► The proposed framework is flexible as it is distribution-free and non-parametric. ► The proposed framework outperforms enumeration and NSGA-II.