Integrating tool acquisition decisions into production planning in flexible manufacturing systems

This research introduces an integrated approach for production planning with tool management in flexible manufacturing systems (FMS). As the most recent development in computerized manufacturing systems, FMSs provide significant competitive advantages in both flexibility and quality. Production planning for such systems involves decision-making over short to intermediate planning horizons. The key decisions are part selection, batching and loading. Proper tooling provides a key source of flexibility for an FMS, ensuring the availability of the right tools at the right machine at the right time in the production of parts. The tooling decision also has direct cost implications. Industry data indicate that tooling accounts for 25-30% of the fixed and variable costs of production in an automated machining environment. While the importance of tool management has been widely recognized, few analytic models in the literature of FMS have integrated tooling considerations into the production planning framework. Specifically, models have not adequately addressed tool costs, the distinction between durable and consumable tools, and effect of the refurbishing delay on tool usage.;This dissertation introduces a mathematical programming-based model integrating tool acquisition and tool allocation decisions into FMS production planning. A two-level hierarchical framework is used to decompose this model into submodels that solve the selection/batching and loading problems separately. The major contribution of this dissertation lies in the new model proposed for solving the FMS selection/batching problem that integrates tool selection and tool cost considerations. The proposed model includes several important tradeoffs in FMS production planning that have not been fully studied in the literature. These tradeoffs are: (1) make or buy (an option to produce less than the full demand), (2) tool cost versus processing time, and (3) tool cost versus setup time.;This dissertation develops a heuristic approach to solve the proposed mixed-integer programming models. The approach developed for solving the selection/batching model uses a simple heuristic, based on linear programming, that solves a sequence of linear programs to successively fix variables at integral values until an all-integer solution is identified. To solve the loading problem, the dissertation proposes a two-stage framework that consists of a simple heuristic, analogues to bin-packing procedures, for allocating tools to machines, and a direct linear programming solution for assigning parts to machines.;The dissertation includes a comprehensive computational evaluation of the planning model and the associated decomposition scheme. To establish a benchmark against the published literature, the computational work also includes a scheme for comparing a specialized version of the proposed model with an existing FMS planning model from the literature.