Facility Location Using Cross Decomposition.
Abstract
Determining the best base stationing for military units can be modeled as a capacitated facility location problem with sole sourcing and multiple resource categories. Computational experience suggests that cross decomposition, a unification of Benders Decomposition and Lagrangean relaxation, is superior to other contemporary methods for solving capacitated facility location problems. Recent research extends cross decomposition to pure integer programming problems with explicit application to capacitated facility location problems with sole sourcing; however, this research offers no computational experience. This thesis implements two cross decomposition algorithm for the capacitated facility location problem with sole sourcing and compares these decomposition algorithms with branch and bound methods. For some problems tested, cross decomposition obtains better solutions in less time; however, cross decomposition does not always perform better than branch and bound due to the time required to obtain the cross decomposition bound that is theoretically superior to other decomposition bounds.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 14, 1995
- Accession Number
- ADA305734
Entities
People
- Leroy A. Jackson
Organizations
- Naval Postgraduate School