Data models for service failure prediction in supply-chain networks

TL;DR

This paper develops data models using machine learning and association rules to predict and explain service failures in last-mile delivery within supply-chain networks, aiming to improve operational efficiency.

Contribution

It introduces a combined approach of supervised classification and association rules for failure prediction and explanation in last-mile delivery, with insights into key failure factors.

Findings

Classifier sensitivity and specificity both around 0.7 for failure types

Confirmation calls significantly reduce customer absence failures

Time window, slack time, and location influence failure likelihood

Abstract

We aim to predict and explain service failures in supply-chain networks, more precisely among last-mile pickup and delivery services to customers. We analyze a dataset of 500,000 services using (1) supervised classification with Random Forests, and (2) Association Rules. Our classifier reaches an average sensitivity of 0.7 and an average specificity of 0.7 for the 5 studied types of failure. Association Rules reassert the importance of confirmation calls to prevent failures due to customers not at home, show the importance of the time window size, slack time, and geographical location of the customer for the other failure types, and highlight the effect of the retailer company on several failure types. To reduce the occurrence of service failures, our data models could be coupled to optimizers, or used to define counter-measures to be taken by human dispatchers.