Optimization of Equipment Replacement: Dynamic Programming-Based Optimization

The purpose of this paper is to present an optimization model formulation based on deterministic dynamic programming (DDP) and to propose both the Bellman and Wagner approaches to solve the problem of equipment replacement optimization (ERO). The developed solution methodology is general and can be used to make optimal keep and replace decisions for new and used vehicles, with and without annual budget considerations. A simple numerical example illustrates and steps through the Bellman DDP solution process and demonstrates how DDP is used to solve the ERO problem through backward recursion. The developed DDP-based ERO software is tested and validated with Texas Department of Transportation vehicle fleet data. Comprehensive numerical results, such as the software computational time and solution quality, are described. Substantial cost savings are estimated with this ERO software. Further research directions are suggested.