Design of maintenance service contracts via principal-agent models

Abstract

In this thesis, the owner of an equipment, called “principal”, aims at outsourcing maintenance services to a specialized service provider (agent). We considered the equipment is subjected to imperfect repairs and several maintenance agents in the market are able to perform the repairs actions. However, these agents are heterogeneous regarding their repair rates (μ). The quickness of the execution of the maintenance task is a private information hold by the agent, which characterizes a scenario of asymmetric information. Given that, we develop a Principal-Agent (PA) model that maximizes principal’s payoff by designing a

contract that entails the best incentives (pricing and penalties) to stimulate agent’s proper self- selection. To model the imperfect repairs, we considered that times to failure are distributed

by a Weibull-Generalized Renewal Process (GRP). Since it is not possible to obtain analytical values of model’s random variables, a Discrete Event Simulation (DES) algorithm is implemented to simulate the agent’s queuing system. To obtain the Nash Equilibrium of the game, the ILOG CPLEX optimization studio is employed to find the optimal solution for a quadratic programming problem. Then, we demonstrated the applicability of the model through an example. Finally, some managerial insights are also provided.