To Trust or to Stockpile: Modeling Human-Simulation Interaction in Supply Chain Shortages

TL;DR

This paper presents a novel modeling approach for human-simulation interaction in supply chain shortages, using PCA, HMM, and sequence analysis to understand decision-making behavior during simulated crises.

Contribution

It introduces a comprehensive method combining PCA, HMM, and sequence analysis to extract behavioral insights from simulation games in supply chain contexts.

Findings

Identified distinct player types: hoarders, reactors, followers.

Behavior varies significantly across different system states.

Sharing information influences decision-making behavior.

Abstract

Understanding decision-making in dynamic and complex settings is a challenge yet essential for preventing, mitigating, and responding to adverse events (e.g., disasters, financial crises). Simulation games have shown promise to advance our understanding of decision-making in such settings. However, an open question remains on how we extract useful information from these games. We contribute an approach to model human-simulation interaction by leveraging existing methods to characterize: (1) system states of dynamic simulation environments (with Principal Component Analysis), (2) behavioral responses from human interaction with simulation (with Hidden Markov Models), and (3) behavioral responses across system states (with Sequence Analysis). We demonstrate this approach with our game simulating drug shortages in a supply chain context. Results from our experimental study with 135 participants show different player types (hoarders, reactors, followers), how behavior changes in different system states, and how sharing information impacts behavior. We discuss how our findings challenge existing literature.