Accelerating Column Generation in Highly Degenerate Integer Programming Problems with Template Pricing

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Marshall, Luke, Shah, Prachi, Dey, Santanu S.
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914470678233088
author Marshall, Luke
Shah, Prachi
Dey, Santanu S.
author_facet Marshall, Luke
Shah, Prachi
Dey, Santanu S.
contents We propose a new pricing strategy for column generation (CG), referred to as Template pricing. This method is motivated by the desire to coordinate solutions of different pricing subproblems in order to accelerate the convergence of the CG process and simultaneously obtain good quality integer feasible solutions. Instead of finding a column with the optimal reduced cost, Template pricing tries to maximize the similarity of columns with a given template vector, while restricting the search to columns with suitable reduced cost. We present an exact and heuristic method (based on Lagrangian relaxation) to efficiently solve the Template pricing problem. We conduct extensive computational experiments on benchmark instances of the Generalized Assignment Problem (GAP). Our results demonstrate that Template pricing can significantly accelerate the CG algorithm, especially in the presence of significant degeneracy, where several benchmark GAP instances solved over 1000x faster than Dantzig pricing, and over 100x with adaptive dual-smoothing. Template pricing allows us to achieve CG optimal bounds on all 1735 ISA instances, finding stronger bounds in 43% and improved integer solutions in 9% of these instances than previously released.
format Preprint
id arxiv_https___arxiv_org_abs_2604_12070
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Accelerating Column Generation in Highly Degenerate Integer Programming Problems with Template Pricing
Marshall, Luke
Shah, Prachi
Dey, Santanu S.
Optimization and Control
90C05
We propose a new pricing strategy for column generation (CG), referred to as Template pricing. This method is motivated by the desire to coordinate solutions of different pricing subproblems in order to accelerate the convergence of the CG process and simultaneously obtain good quality integer feasible solutions. Instead of finding a column with the optimal reduced cost, Template pricing tries to maximize the similarity of columns with a given template vector, while restricting the search to columns with suitable reduced cost. We present an exact and heuristic method (based on Lagrangian relaxation) to efficiently solve the Template pricing problem. We conduct extensive computational experiments on benchmark instances of the Generalized Assignment Problem (GAP). Our results demonstrate that Template pricing can significantly accelerate the CG algorithm, especially in the presence of significant degeneracy, where several benchmark GAP instances solved over 1000x faster than Dantzig pricing, and over 100x with adaptive dual-smoothing. Template pricing allows us to achieve CG optimal bounds on all 1735 ISA instances, finding stronger bounds in 43% and improved integer solutions in 9% of these instances than previously released.
title Accelerating Column Generation in Highly Degenerate Integer Programming Problems with Template Pricing
topic Optimization and Control
90C05
url https://arxiv.org/abs/2604.12070