Accelerating Dynamic Programming via Dual Selection for Column-Generation-Based Exact Solution Frameworks

This talk examined how dual selection can accelerate dynamic programming components inside column-generation-based exact solution frameworks. The central idea is to reduce avoidable computational burden in pricing-style subproblems while preserving the exactness and structural advantages of the larger decomposition framework.

The presentation was supported by travel funding from the 2024 INFORMS Optimization Society Conference and reflected my broader interest in making exact optimization methods more scalable in practice.

• Venue: 2024 INFORMS Optimization Society Conference
• Related paper: coming soon
• Slides PDF: coming soon