A new formulation for the collection and delivery problem of biomedical specimen

TL;DR

This paper introduces a new mixed-integer programming formulation for the biomedical specimen collection and delivery problem, enabling faster exact solutions for larger instances than previous methods.

Contribution

It proposes a novel two-index formulation that improves solution efficiency and scalability for the CDSP in healthcare logistics.

Findings

Outperforms existing models and metaheuristics

Solves 80 out of 168 benchmark instances to optimality

Handles instances with up to 100 delivery requests

Abstract

We study the collection and delivery problem of biomedical specimens (CDSP) with multiple trips, time windows, a homogeneous fleet, and the objective of minimizing total completion time of delivery requests. This is a prominent problem in healthcare logistics, where specimens (blood, plasma, urin etc.) collected from patients in doctor's offices and hospitals are transported to a central laboratory for advanced analysis. To the best of our knowledge, available exact solution approaches for CDSP have been able to solve only small instances with up to 9 delivery requests. In this paper, we propose a two-index mixed-integer programming formulation that, when used with an off-the-shelf solver, results in a fast exact solution approach. Computational experiments on a benchmark data set confirm that the proposed formulation outperforms both the state-of-the-art model and the state-of-the-art metaheuristic from the literature, solving 80 out of 168 benchmark instances to optimality, including a significant number of instances with 100 delivery requests.