Integrare la Controllabilità nella Valutazione Quantitativa del Rischio di Programmazione: uno Studio basato su Simulazioni Monte Carlo

Abstract

Understanding how uncertainty propagates through project networks requires tools that capture how distributional assumptions shape schedule risk behavior. This thesis develops a quantitative framework that analyzes activity‑level sensitivity by representing controllability through the width of activity‑duration distributions and expertise through distributional shape and mode position. Twelve scenarios, combining three uncertainty ranges with four distribution groups (symmetric, right‑skewed, left‑skewed, and uniform), are applied to representative activities in project networks. For each scenario, Monte Carlo simulations generate project completion times, from which Schedule Sensitivity Index (SSI) values are computed and examined using mean rankings, clustering of sensitivity patterns, mode‑position effects, and variance‑based stability metrics at both single‑network and multi‑project levels. ​ The results reveal a stable subset of structurally influential activities whose high SSI values persist across all scenarios, alongside a second group of controllable activities whose SSI responds strongly to changes in uncertainty range and distributional shape. These controllable activities exhibit sensitivity profiles that can increase or decrease under different uncertainty conditions, making them primary targets for managerial intervention. Left‑skewed and uniform distributions consistently raise SSI levels and dispersion, whereas narrower ranges produce more stable but less adjustable sensitivity patterns than wider ranges. ​ Overall, the study operationalizes controllability as a quantitative property, clarifies its interaction with distributional assumptions, and offers a structured basis for knowledge‑aware project risk prioritization and future methodological development.

Understanding how uncertainty propagates through project networks requires tools that capture how distributional assumptions shape schedule risk behavior. This thesis develops a quantitative framework that analyzes activity‑level sensitivity by representing controllability through the width of activity‑duration distributions and expertise through distributional shape and mode position. Twelve scenarios, combining three uncertainty ranges with four distribution groups (symmetric, right‑skewed, left‑skewed, and uniform), are applied to representative activities in project networks. For each scenario, Monte Carlo simulations generate project completion times, from which Schedule Sensitivity Index (SSI) values are computed and examined using mean rankings, clustering of sensitivity patterns, mode‑position effects, and variance‑based stability metrics at both single‑network and multi‑project levels. ​ The results reveal a stable subset of structurally influential activities whose high SSI values persist across all scenarios, alongside a second group of controllable activities whose SSI responds strongly to changes in uncertainty range and distributional shape. These controllable activities exhibit sensitivity profiles that can increase or decrease under different uncertainty conditions, making them primary targets for managerial intervention. Left‑skewed and uniform distributions consistently raise SSI levels and dispersion, whereas narrower ranges produce more stable but less adjustable sensitivity patterns than wider ranges. ​ Overall, the study operationalizes controllability as a quantitative property, clarifies its interaction with distributional assumptions, and offers a structured basis for knowledge‑aware project risk prioritization and future methodological development.