|dc.description.abstract||This dissertation is based on three essays in operations and supply chain management.
In essay 1, we study an operations scheduling problem in a complex manufacturing system, most notably, semiconductor manufacturing. In particular, we study the scheduling problem of minimizing total weighted tardiness on parallel non-identical batch processing machines. We formulate the (primal) problem as a nonlinear integer programming model. Moreover, we prove that the primal problem can be solved exactly by solving a corresponding dual problem with nonlinear relaxation. Since both the primal and the dual problems are NP-hard, we propose to use genetic algorithms, based on random keys and multiple choice encodings, to heuristically solve them. We found that the genetic algorithms consistently outperform a standard mathematical programming package in terms of solutions and computation times. We also found that for small scale problem instances, the multiple choice genetic algorithm outperforms the random keys genetic algorithm, while for medium and large scale problem instances, the random keys genetic algorithm outperforms the multiple choice genetic algorithm.
In essay 2, we study a monopolist firm offering successive versions of a durable good (e.g., software) that improves over time. The firm decides the time between successive introductions as well as price. In turn, consumers strategically decide whether to purchase or wait for a later version. We model and analyze three alternative strategies for offering successive product versions: the partial-, continuous-, and no-updates policies. We first consider the firm's profit maximizing policy assuming a homogeneous market and subsequently address consumers with heterogeneous product valuations. Our analytic model's simple structure and results highlight the important tradeoff between price and release timing for products with successive versions.
In essay 3, we study the effect of time series structure of customer demand models on the value of information sharing within a supply chain. We contribute to the literature by incorporating a nonlinear demand model based on exponential disturbances, coupled with temporal heteroscedasticity, which captures more complex patterns in the demand process. We examine the conditions under which information sharing is valuable.||en_US