Solution Strategies for the Design and Planning of Supply Chains and Embedded Plants

In this work we present solution strategies for the task of designing supply chains with the explicit consideration of the detailed plant performance of the embedded facilities. Taking as a basis an mixed integer linear programming (MILP) model introduced in a previous work, we propose three solution strategies that exploit the underlying mathematical structure. The strategies are based on bi-level decomposition algorithm and Lagrangean decomposition method. Also, we propose a hybrid approach that takes advantage of both well known strategies, improving the solution obtained by Lagrangean methods. Numerical results show that the bi-level approach works more efficiently than Lagrangean and hybrid decomposition, and lead to significant CPU savings when compared to the full space MILP for large scale problems.