E-Commerce Research Paper: Dynamic Forecast Demand Analysis to Design a Parcel Locker Delivery Network in Spain

Introduction

Business logistics faced significant challenges during the COVID-19 pandemic due to disruptions in last-mile delivery processes. Self-collection delivery systems and automated parcel lockers emerged as promising solutions, offering flexibility, reducing delivery times, and minimizing gas emissions.

In this e-commerce research paper, you'll learn how simulation modeling helps plan parcel locker delivery networks, considering demand.

Solution

The 2023 study extends the “A Simulation-Optimization Model for Automated Parcel Lockers Network Design” research. The follow-up paper uses the hybrid model introduced in the previous material. It combines an agent-based approach with a facility location problem (FLP) for parcel locker delivery network design.

In the latest e-commerce research paper, the scholars added demand forecasting by means of the system dynamics method. The objective was to minimize the number of automated parcel lockers while maximizing demand coverage.

Scheme of the simulation-optimization framework design

Flowchart of the simulation-optimization framework

The methodology involves:

  1. System dynamics simulation model: The model forecasts e-commerce demand in Pamplona (Spain) by analyzing socioeconomic factors such as population growth, internet usage, and shoppers' behaviors. It captures how potential e-shoppers and users of parcel delivery lockers interact and influence demand patterns.
  2. ε-Constraint: The method solves a bi-criteria facility location problem. It balances the number of parcel delivery lockers and their coverage of e-commerce demand. The model helps to find the best locations for automated parcel lockers by considering constraints like the maximum travel distance for customers and the capacity of each parcel delivery locker. It ensures efficient placement and good service coverage.
  3. Integration and tools: The simulation model is built in AnyLogic software, and optimization is managed with IBM ILOG CPLEX. This integrated framework dynamically adjusts the parcel locker delivery network to align with evolving demand forecasts, ensuring the system remains responsive and efficient.

Results

The simulation experiments presented in this e-commerce research paper tested demand coverage scenarios (25%, 50%, 75%, and 90%), showing that higher coverage requires more automated parcel lockers, particularly in the city outskirts.

The model effectively balances the network design and demand coverage of parcel delivery lockers, reducing delivery times and environmental impact. The network’s scalability suggests it can be applied to other European cities. The study presents a robust framework combining system dynamics for forecasting and bi-criteria facility location optimization.

相关文章