Stochastic Optimization Problems in the Service Industry with Customer Considerations

This dissertation studies the effect of customer considerations in stochastic optimization problems in the service industry. The first part looks at problems in the service industry geared towards providing efficient service, i.e., minimizing the total cost of providing service while ensuring that certain service level targets are met. In particular, we study applications in the call center industry, repair facilities, and spare parts industry. We develop mathematical programming models for these problems to determine the optimal non-randomized policy, i.e., a policy not involving any probabilistic mixing. The service level targets in these problems can be naturally modeled as constraints, which can be incorporated into a mathematical program. Despite this fact, to the best of our knowledge, this dissertation is the first study to use such a formulation to find an optimal non-randomized policy. The second part of this dissertation considers a revenue-management setting in the service industry but, different from the existing literature, the seller has to negotiate the price with each buyer. This is a typical situation in B2B applications such as telecommunications, cargo shipping, etc; in such settings the seller negotiates a different sale price with different buyers. We also consider a model in which the seller ensures that the average buyer is no worse-off by bargaining than under posted-pricing, via a constraint. We characterize how the seller and the average buyer are effected. We obtain several additional insights through numerical experiments.