LAUSR

Abstract Stochastic time-cost-quality tradeoff problem (STCQTP) analysis significantly expands the scope of discrete (deterministic) project duration-cost analysis. STCQTP requires multi-objective optimization methodologies to locate optimum total-project-cost-quality solutions for facilities-construction projects with multiple parallel pathways of work items involving high degrees of duration, cost and quality uncertainties; a situation common in the gas and oil industry. Calculating Pareto frontiers of non-dominated-total-project-cost and total-project-quality solutions across the range of feasible project durations further extends the usefulness of STCQTP analysis. For stochastic analysis project-work-item durations and costs are expressed as probability distributions and sampled with random numbers (0,1). By controlling the fractional numbers used to sample the work-item cost distributions by formulas linked to the random numbers used to sample the work-item duration distribution, a wide range of complex time-cost relationships can be defined. Fuzzy analysis is applied to each stochastic case generated to integrate more subjective assessments of work-item quality achieved. A memetic algorithm, developed for constrained STCQTP involves ten metaheuristics configured to combine local exploitation and global exploration of the feasible duration-cost solution space. Fuzzy analysis of work-item quality is integrated with each stochastic scenario evaluated. The proposed algorithm effectively delivers realistic multiple objectives of: 1) global total-project-cost minima; 2) global total-project-quality minima; and, 3) Pareto frontiers of non-dominated total-project duration versus cost, duration versus quality, and/or duration versus total-project-cost-quality function test score. Analysis of an example gas-processing-plant-construction project, applying three distinct work-item duration-cost relationships, demonstrates with the aid of metaheuristic profiling, that the memetic STCQTP algorithm coded in visual basic for applications for execution via an Excel spreadsheet, requires no proprietary software to deliver its objectives. Dynamic adjustment factors applied by some metaheuristics, derived from fat-tailed distributions sampled by chaotic sequences, aid efficient searching of the feasible solution space. The metaheuristic profiles also help to fine tune the metaheuristic configurations of the algorithm applied to specific project cases.