Corner cutting in complex engineering systems: a game theoretic and probabilistic modeling approach

Corner cutting in complex engineering systems is a significant problem because a departure from project specifications can be the cause of costly system failure. Covert corner cutting by an agent exacerbates the problem because the agent is imposing risks upon the system that management is not aware of and that they would not necessarily accept. The challenge for management (the principal) is to understand how their policies toward the agent (payments, penalties, and inspection) affect the agent's actions, and, in turn, how the agent's actions affect the system's performance. The resulting framework guides management in the setting of their optimal policies for numerous situations, including when the principal prefers illicit agent corner cutting, when the principal is unable to stop the agent from corner cutting, and when the agent is naive and does not react to the principal's incentives. This dissertation links probabilistic risk analysis and engineering economics with expected-utility decision analysis and the principal-agent framework to create a decision support system for upper management to aid in the optimal setting of the agent's incentives. Probabilistic risk analysis is needed to assess the effect of corner cutting on system performance. Engineering economics converts the effects of corner cutting to its monetary equivalent through cash-flow analysis. Expected-utility decision analysis is used to determine both the principal's and the agent's optimal actions under various conditions, and the principal-agent framework models the interaction between the principal's parameter settings and the agent's decisions and actions.