This document outlines the essential elements for a simulation environment designed to study hybrid last-mile delivery systems, combining human couriers and Autonomous Delivery Robots (ADRs).

1. Road Network:

   • Detailed road network data from OpenStreetMap (OSM).
   • Accurate representation of streets, intersections, sidewalks, and bike lanes.
   • Road attributes like speed limits, traffic directions, and road types.

2. Buildings and Obstacles:

   • Building footprints and heights for realistic 3D representation.
   • Obstacles like trees, streetlights, and parked cars.

3. Traffic:

   • Dynamic traffic simulation with varying densities and flow patterns.
   • Different vehicle types (cars, buses, trucks, bicycles) with realistic behaviors.
   • Simulation of traffic signals and pedestrian crossings.

4. Delivery Locations:

   • Realistic distribution of customer locations (residential, commercial).
   • Clustering of delivery locations to simulate demand hotspots.

1. Human Couriers:

   • Different courier types (on foot, bicycle, motorcycle, van).
   • Varying speeds, capacities, and operating costs.
   • Realistic behaviors like obeying traffic rules and navigating sidewalks.

2. Trucks:

   • Traditional delivery trucks for potential replenishment or as mobile hubs for ADRs.
   • Varying sizes and capacities.

3. Autonomous Delivery Robots (ADRs):

   • Different types of ADRs (sidewalk, road-based) with varying speeds, ranges, and capacities.
   • Navigation algorithms for path planning and obstacle avoidance.
   • Battery life and charging behavior.

1. Order Generation:

   • Dynamic order generation with varying origins, destinations, and time windows.
   • Order priorities and types (e.g., size, weight, fragility).

2. Task Allocation:

   • Dynamic allocation of orders to human couriers or ADRs based on real-time conditions and agent capabilities.

3. Dispatching and Routing:

   • Algorithms for efficient routing and scheduling of both human couriers and ADRs.
   • Consideration of traffic conditions, delivery time windows, and agent constraints.

1. Demand Forecasting:

   • AI model (e.g., LSTM) for predicting future demand patterns.

2. AI Solver (Optional):

   • Graph-based neural network or other AI approach for optimizing routing and delivery decisions.

1. Data Collection:

   • Collect data on delivery times, distances, energy consumption, and other relevant metrics.

2. Performance Evaluation:

   • Use metrics like average delivery time, total cost, success rate, and customer satisfaction to evaluate the efficiency of the hybrid system.