Solving the geometric firefighter routing problem via integer programming

•New GFP variants that allow for more general sets of barriers and solutions.•Problems solved to optimality through Integer Programming modeling.•Preprocessing algorithms and IP-based primal heuristics described in detail.•Two large publicly available sets of challenging instances based on actual fo...

Full description

Saved in:
Bibliographic Details
Published in:European journal of operational research Vol. 274; no. 3; pp. 1090 - 1101
Main Authors: O. Zambon, Mauricio J., J. de Rezende, Pedro, C. de Souza, Cid
Format: Journal Article
Language:English
Published: Elsevier B.V 01.05.2019
Subjects:
Combinatorial optimization
Computational geometry
Geometric firefighter routing problem
Integer programming
Integer programming
Computational geometry
Combinatorial optimization
Geometric firefighter routing problem
ISSN:0377-2217, 1872-6860
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:•New GFP variants that allow for more general sets of barriers and solutions.•Problems solved to optimality through Integer Programming modeling.•Preprocessing algorithms and IP-based primal heuristics described in detail.•Two large publicly available sets of challenging instances based on actual forests.•Openly accessible best known bounds and solutions for more than a thousand instances. In this paper, we introduce the Geometric Firefighter Routing Problem (gfrp) as a variant of the Geometric Firefighter Problem aiming to better model more realistic situations. We design an exact algorithm based on a core Linear Integer Programming formulation and propose additional sets of valid constraints to strengthen it. The algorithm also includes primal heuristics, and preprocessing procedures to reduce the model size. Besides, we generate two large sets of instances, tailored to the gfrp, and report on comprehensive experimental results for them. Thorough analysis validate the effectiveness of each major step of the algorithm and the overall performance of our approach.
ISSN:0377-2217
1872-6860
DOI:10.1016/j.ejor.2018.10.037