A simulation environment for multiple project resource optimization

Abstract

This thesis develops a numerical simulation environment as a management support tool applicable to the selection and scheduling of multiple, concurrent research and development projects under conditions of constrained resources and uncertain program requirements. A prototype version of this software tool, called SEMPRO (Simulation Environment for Multiple Project Resource Optimization), is developed to capture an operational model of the NASA Dryden Flight Research Center flight and a representative research project portfolio. An attribute-driven Work Breakdown Schedule generates resource-loaded activity networks for each entry in the research project portfolio. The project selection and project coordination problems are formulated as Binary Integer Linear Programming problems, as extensions of the traditional Resource Constrained Project Scheduling Problem (RCPSP). To alleviate the computational obstacles associated with these NP Hard problems, a Lagrange Relaxation formulation is used to generate a near-optimal, time-phased sequence for execution of the selected project activity networks. Stochastic, non-linear, discrete-event simulation dynamics are then employed to validate these linear optimization solutions against a representative model of the organization's research and development project operational processes. The SEMPRO prototype is written in the Microsoft Excel Visual Basic Application language to facilitate project management visualization and knowledge transfer.