Urban drone stations siting optimization based on hybrid algorithm of MILP and machine learning.

Area of Science:

• Operations Research
• Urban Planning
• Emergency Management

Background:

• Urban areas face critical emergencies like fires and accidents due to high density and infrastructure.
• Inefficient distribution of emergency facilities leads to prolonged response times.
• Existing emergency service locations in Chengdu were found to be suboptimal.

Purpose of the Study:

• To identify optimal locations and numbers for Unmanned Aerial Vehicle (UAV) fire stations and drone ambulance centers.
• To improve emergency response times in urban environments.
• To develop a robust framework for siting and allocating emergency service facilities.

Main Methods:

• A two-stage clustering method (X-means and K-means) was used to determine optimal facility placement.
• A Mixed-Integer Linear Programming (MILP) model was developed and solved using the Gurobi platform.
• Bayesian optimization (machine learning) was employed to analyze response speed and service capacity.

Main Results:

• The study identified inefficiencies in current emergency facility distribution, leading to suboptimal response times.
• The hybrid MILP and machine learning approach provided a robust solution for facility siting and allocation.
• The optimized layout demonstrated potential for enhanced emergency preparedness and faster response in urban settings.

Conclusions:

• Integrating Mixed-Integer Linear Programming and machine learning offers a powerful framework for complex emergency service facility problems.
• The proposed hybrid algorithm significantly enhances urban emergency preparedness and response capabilities.
• The use of Unmanned Aerial Vehicle (UAV) stations presents a viable strategy for improving emergency service delivery.