Neurologistics Simulation Code: Neuro-Inspired Forecasting of Ethical Decision-Making in Sustainable Supply Chains
Description
This dataset contains the full Python codebase used in the study "Neurologistics: A Neuro-Inspired Simulation for Forecasting Ethical Decision-Making in Sustainable Supply Chains." The Neurologistics framework models how biologically inspired cognitive traits—empathy, stress tolerance, and impulse control—influence ethical trade-offs in logistics operations under pressure from emerging technologies (e.g., AI optimization, autonomous delivery, ESG dashboards). The simulation produces 15,000 agent-based decisions across three ethical scenarios: speed vs. emissions, cost vs. labor rights, and efficiency vs. sustainability. It includes modules for simulation execution, econometric analysis, structural equation modeling, robustness checks, and visualizations. Several files simulate learning mechanisms and compare agent behavior across time, enabling the study of ethical adaptation. The code supports generating figures in the manuscript and allows replication of all reported findings.
Files
Steps to reproduce
This dataset accompanies the manuscript "Neurologistics: A Neuro-Inspired Simulation for Forecasting Ethical Decision-Making in Sustainable Supply Chains." It contains Python simulation and analysis scripts that allow researchers to replicate the study's main results. To reproduce the findings, users should begin by running the simulation model, which generates synthetic data representing agent-based ethical decisions under varying levels of empathy, stress, and impulse control. The simulation mimics moral dilemmas in logistics operations - trade-offs between delivery speed and carbon emissions or cost efficiency and labor rights. Following data generation, separate scripts are provided to conduct econometric analyses (logistic regression, OLS, and structural equation modeling) on the simulated outcomes. These analyses quantify how neuro-inspired cognitive traits predict ethical versus efficiency-oriented decisions. A series of visualization scripts produce the core figures from the paper, including bar plots of ethical decisions by scenario, ethics vs. efficiency scatterplots, and a correlation heatmap of cognitive traits. The dataset includes robustness checks and sensitivity analyses, demonstrating the stability of model results when the parameters of empathy or stress tolerance are varied. Other scripts explore how learning over time affects ethical decision patterns, simulating organizational adaptation and ethical growth in response to feedback. Users are encouraged to execute the scripts sequentially to replicate the results, beginning with the simulation, followed by model estimation, visualization, and learning effects. All scripts are organized and can be run independently once the simulation data is available. This modular structure enables transparent replication, extension of the model, or adaptation to new use cases in ethical AI, logistics, and behavioral forecasting.
Institutions
- Universitatea de Vest din Timisoara Facultatea de Economie si de Administrare a Afacerilor