A Lasserre Lower Bound for the Min-Sum Single Machine Scheduling Problem

TL;DR

This paper investigates the limitations of the Lasserre hierarchy in approximating the Min-sum single machine scheduling problem, showing that the integrality gap remains unbounded at certain levels even for simple variants.

Contribution

It provides the first lower bound for the Lasserre hierarchy on this problem, demonstrating unbounded integrality gaps at level ( ) for a problem solvable efficiently by classical algorithms.

Findings

Lasserre hierarchy has unbounded integrality gap at level ( a)

The problem variant studied is solvable in O(n log n) time by Moore-Hodgson algorithm

Partial diagonalization technique is used to prove the lower bound.

Abstract

The Min-sum single machine scheduling problem (denoted 1||sum f_j) generalizes a large number of sequencing problems. The first constant approximation guarantees have been obtained only recently and are based on natural time-indexed LP relaxations strengthened with the so called Knapsack-Cover inequalities (see Bansal and Pruhs, Cheung and Shmoys and the recent 4+\epsilon-approximation by Mestre and Verschae). These relaxations have an integrality gap of 2, since the Min-knapsack problem is a special case. No APX-hardness result is known and it is still conceivable that there exists a PTAS. Interestingly, the Lasserre hierarchy relaxation, when the objective function is incorporated as a constraint, reduces the integrality gap for the Min-knapsack problem to 1+\epsilon. In this paper we study the complexity of the Min-sum single machine scheduling problem under algorithms from the Lasserre hierarchy. We prove the first lower bound for this model by showing that the integrality gap is unbounded at level \Omega(\sqrt{n}) even for a variant of the problem that is solvable in O(n log n) time by the Moore-Hodgson algorithm, namely Min-number of tardy jobs. We consider a natural formulation that incorporates the objective function as a constraint and prove the result by partially diagonalizing the matrix associated with the relaxation and exploiting this characterization.