A reliability-aware randomized simheuristic for the stochastic team orienteering problem

TL;DR

This paper introduces a reliability-aware simheuristic for the stochastic team orienteering problem with lognormal travel times, combining heuristic strategies to improve solution reliability and computational efficiency.

Contribution

It proposes a novel simheuristic integrating a Top-L randomization, stochastic screening, and reliability thresholds, offering a simpler yet effective alternative to existing methods.

Findings

The method is competitive with VNS-based approaches on certain instances.

It shows significant gains on two-vehicle long-route instances.

The simpler architecture outperforms on smaller instances, highlighting a trade-off.

Abstract

We study a stochastic variant of the Team Orienteering Problem with lognormal travel times and an all-or-nothing reward policy, under which the reward of a route is lost if its travel time exceeds the available budget. We propose a reliability-aware simheuristic that combines a savings-based constructive heuristic with three specific design elements: a Top-$L_{\mathrm{top}}$ randomization mechanism, stochastic screening of the savings parameter, and an explicit reliability threshold for solution selection. Computational experiments on the Chao et al. benchmark show that the method is competitive with the VNS-based simheuristic of Panadero et al. (2020) on a non-trivial subset of instances using a significantly simpler architecture, with the largest gains on two-vehicle sub-families with long routes where the reliability-aware selection compensates for the absence of VNS-style exploration. On larger multi-vehicle instances the simpler architecture is outperformed, and this trade-off is discussed explicitly.