High-level integrated deterministic, stochastic and fuzzy cost-duration analysis aids project planning and monitoring, focusing on uncertainties and earned value metrics

Abstract

Integrating deterministic, fuzzy and stochastic analysis of cost-duration progress of complex projects under varying conditions of uncertainty at a high-level (at the work-item rather than the individual activity level) can be beneficial for decision makers in planning and monitoring infrastructure and other complex projects. Incorporating various facets of cost-duration uncertainty analysis with the principles of earned value management (EVM) can provide significant insight to the budget and schedule performance of projects with multiple parallel pathways of work items, plus reliable to-completion forecasts as a project evolves. Focusing on the critical path, stochastic analysis is able to quantify criticality, cruciality, uncertainty and downside risk measures at project, work item and budget levels. A project network and critical path analysis built around work breakdown progress diagrams calculating the progress to completion of between 20 and 50 work items at regular intervals (e.g. 2%–5%, involving 50 to 20 points equally spaced in time) along a baseline planned project schedule, provides a useful framework for a high-level cost-duration model. That framework can be rapidly and consistently evaluated for each case selected applying deterministic, fuzzy and stochastic analysis, each providing complementary insight to a project's performance at specific points in time, to-completion cost-duration forecasts, and quantify downside risks and uncertainties on a range of budget and schedule targets. A methodology is proposed that calculates earned duration and related duration performance index for critical path items weighted for their planned durations provides a measure of project duration performance that is more focused on critical path and crucial work items than standard earned schedule and earned duration metrics. Fuzzy analysis associated with the inability to establish precisely what progress has been truly achieved on each work items adds an additional component to uncertainty analysis not provided by stochastic analysis. Through careful selection of fuzzy set definitions and defuzzification methods fuzzy and stochastic models can be tuned to provide comparable and reliable EVM performance measures and improved to-completion forecasts with low mean absolute percentage errors to actual outcomes. The proposed methodology provides decision makers with a flexible and easy-to-interpret analysis, on a scale that is easy to produce in VBA driven spreadsheet without recourse to proprietary software, integrating multiple perspectives of uncertainty.