Plant location and capacity planning for components procurement under global manufacturing environments

Abstract

This dissertation focuses on plant location and capacity planning for components procurement under global manufacturing environments. For the economical procurement and/or supply for components needed by globally dispersed plants (customers), the schedule of opening components manufacturing plants, plans for acquisition of capacities in opened components manufacturing plants, and plans for components procurement are determined. Several practical considerations in plant location and capacity planning for components manufacturing and procurement are considered such as: types, sizes and functional characteristics of facilities; limitations on investments for opening plants and purchasing facilities; a special requirement for components procurement in global manufacturing environments. First, a problem of multiperiod capacity planning for components manufacturing plants is solved to determine a mix of flexible and dedicated capacities under budget restriction. Schedules of acquiring and replacing facilities are determined and operations are assigned to the flexible and dedicated facilities. The problem is formulated as a mixed integer linear program and solved by the Lagrangian relaxation approach. A subgradient optimization method and a multiplier adjustment method are employed to obtain better lower bounds. An LP-based Lagrangian heuristic algorithm is developed to find good feasible solutions. Second, a dynamic plant location and capacity planning problem is solved for distributed components manufacturing and procurement/supply. The opening schedule of components manufacturing plants, allocations of customers```` demands and plans for acquisition and/or disposal of plants capacities are determined. The problem is formulated as a mixed integer linear program and solved by a heuristic algorithm based on Lagrangian relaxation and a cut and branch algorithm which uses Gomory cuts. Several solution properties of the relaxed problem are found and used to develop effic...