|
Original research INTEGRATED PROJECT PORTFOLIO SELECTION AND SCHEDULING UNDER FUZZY-STOCHASTIC UNCERTAINTY: A MULTI-AGENT MULTI-OBJECTIVE OPTIMIZATION APPROACHPages 431-446 Abstract
This study proposes a multi-objective fuzzy–stochastic programming model for simultaneous project portfolio selection, scheduling, and allocation within a multi-agent organizational structure, integrating economic, environmental, and resilience objectives. Fuzzy numbers represent uncertainty in project construction and execution times, while stochastic scenarios capture variability in future conditions. Small instances were evaluated using the ε-Constraint method and BARON in GAMS, whereas larger instances were solved using NSGA-II and MOPSO. NSGA-II closely matched exact reference solutions for small instances and handled problems with up to 40 projects, 8 agents, and 4 scenarios. MOPSO achieved better MID values, whereas NSGA-II performed better in the number of Pareto solutions, crowding distance, and computational time. Sensitivity analysis showed that higher discount rates reduced economic value by 32.6%, while higher reinvestment rates increased it by up to 86.9%. The model supports integrated portfolio decisions under uncertainty.
Keywords:
Project portfolio selection and scheduling, multi-objective programming, fuzzy–stochastic model, multi-agent structure,
economic and environmental resilience, NSGA-II.
|
Hadis Gholami, 
PDF 