| Semiconductor manufacturers often have a limited capacity to make integrated-circuits since capacity expansion is very costly and time consuming. Thus, when demand exceeds supply, the manufacturer, as a component supplier, needs to put customers on allocation. At the strategic level, game theoretic models are used in this research to analyze strategic behaviors of the supplier and the buyers in an allocation game with Cournot competition among the buyers. While a preference-based policy can minimize the supplier's total shortage penalty, conditions under which different capacity allocation policies work better for not only the supplier's own profitability but also the total supply chain performance are derived.;At the operational level, the Generalized Lot Allocation problem of how to configure and allocate semiconductor product lots to customer orders under the preference-based policy is modeled as a combinatorial optimization problem with the objective of minimizing total shortage and tardiness penalties as well as total product downgrade and product excess costs. The composite nonlinear integer programming formulation of the problem can be equivalently transformed into a more tractable linear mixed integer program. However, the lot-integrity requirement, a de facto standard in the semiconductor industry, makes the problem computationally intractable. Thus, a set of heuristic-based lot allocation methods is developed and compared in this research. Computational test results illustrate that our proposed decision procedures can outperform current practice consistently, while the composite allocation rule-based methods can help keep a proper balance between the quality of the solution and the cost of the solution. |
