Methods and Benefits of Integrating the Virtual Design Team with Probabilistic Risk Analysis for Design Project and Program Planning

This study focuses on complex but routine design, in accordance with VDT's principal experience base. I propose methods by which the Virtual Design Team (VDT) project simulator and Probabilistic Risk Analysis (PRA) technologies may be deeply synthesized to support design project and program planning calculations such as optimal strategy and policy, value of information, and marginal value of resources. VDT and PRA are theoretically and methodologically compatible, and pragmatically complementary in several important ways. PRA has been extended to include human and organizational elements while addressing interdependencies among multiple projects' decisions, but it does not offer substantial guidance on central, challenging problems including project schedule and cost estimation. In practice, the complexity of project behavior routinely eludes expert opinion, the typical PRA data source. VDT offers to compensate for this deficiency by breaking the complexity into manageable and theoretically grounded elements. Conversely, VDT predicts many subtle and important measures of risk, but does not offer sufficient comprehensiveness or precision to address several important programmatic risks. The VDT system also does not model strategic mid-project decisions, and does not directly recommend action, but PRA can provide these functions. In this paper, I describe three methods of integrating PRA and VDT to take advantage of these various complementarities in a practical setting. I explain the key contributions that each of the integration models offers over standalone VDT and PRA, and then detail a combined formulation that employs each of the three integrations to address programs that include multiple, dependent projects. Currently, the proposed tools are particularly well justified for design projects that culminate in a risky endeavor such as a space mission. The paper concludes with an algorithmic complexity assessment, model justification plan, and sample reports.