Interdicting a Force Deployment: Two-Sided Optimization of Asset Selection, Lift Scheduling, and Multi-Commodity Load Planning
Abstract
A military deployment is visible and vulnerable. But, deployments are currently planned assuming they can be completed with surprise, or defended from any threat. JFAST, the current deployment planning and visualization tool of choice, uses heuristics of unknown reliability that yield deployment plans of unknown quality, and ignores vulnerability. We introduce LIFTER, an integer-linear program (ILP) that optimizes a time-phased force deployment (TPFDD) by day, by asset cycle, and by TPFDD line (individual shipment from an origin to a destination), and ATTACKER, also an ILP, representing a smart enemy's resource-limited interdictions to maximally disrupt LIFTER's subsequently re-optimized TPFDD plan. LIFTER activates transport assets from an allocation list, and yields a complete logistic plan that minimizes disruption represented by penalties for early, tardy, late, or dropped shipments, and for under-utilization of asset capacity. We use LIFTER to qualitatively assess JFAST heuristic plans. We also link both ILPs in a decomposition-based search for the best deployment plan around the worst-case interdiction, given that the actions of deployer and interdictor are transparent to both parties. We explain how JFAST could be embellished with its own version of ATTACKER. A key discovery here is a gauge of the value of intelligence, deception, and secrecy.
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 01, 2005
- Accession Number
- ADA432887
Entities
People
- Peter M. Koprowski
Organizations
- Naval Postgraduate School