Robust transshipment problem under consistent flow constraints

TL;DR

This paper investigates the computational complexity of a robust transshipment problem with flow constraints, demonstrating NP-hardness results, and providing polynomial algorithms for specific graph classes and scenario counts.

Contribution

It establishes NP-hardness of the problem on acyclic and series-parallel digraphs, and offers polynomial algorithms for certain special cases and graph types.

Findings

NP-hard on acyclic digraphs

Weakly NP-hard on series-parallel digraphs

Polynomial algorithms for specific cases and pearl digraphs

Abstract

In this paper, we study robust transshipment under consistent flow constraints. We consider demand uncertainty represented by a finite set of scenarios and characterize a subset of arcs as so-called fixed arcs. In each scenario, we require an integral flow that satisfies the respective flow balance constraints. In addition, on each fixed arc, we require equal flow for all scenarios. The objective is to minimize the maximum cost occurring among all scenarios. We show that the problem is strongly NP-hard on acyclic digraphs by a reduction from the $(3,B2)$-SAT problem. Furthermore, we prove that the problem is weakly NP-hard on series-parallel digraphs by a reduction from a special case of the \textsc{Partition} problem. If in addition the number of scenarios is constant, we observe the pseudo-polynomial-time solvability of the problem. We provide polynomial-time algorithms for three special cases on series-parallel digraphs. Finally, we present a polynomial-time algorithm for pearl digraphs.