Multi-objective routing optimization for hazardous material vehicles integrating Industry 4.0 technologies, heterogeneous fleets, time windows, and time-dependent networks using NSGA-III and multi-objective metaheuristic algorithms

Multi-objective routing optimization for hazardous material vehicles integrating Industry 4.0 technologies, heterogeneous fleets, time windows, and time-dependent networks using NSGA-III and multi-objective metaheuristic algorithms

Farshid Amireinanloo1 Mehran Khalaj2 Davood Jafari3

1) Department of Industrial Engineering Pa.C., Islamic Azad University, Parand, Iran
2) Department of Industrial Engineering Pa.C., Islamic Azad University, Parand, Iran
3) Department of Industrial Engineering Pa.C., Islamic Azad University, Parand, Iran

IOI Code : ABEB-ACZGEK
Publication : 4th International Conference on Recent Advances in Engineering, Innovation and Technology - Belgium(eitconf.com)
Abstract :
    Routing hazardous materials entails simultaneous decision-making regarding transportation costs, accident probability and consequences, pollutant emissions, and service reliability. This paper presents an integrated framework for a multi-objective routing problem featuring a heterogeneous fleet, hard time windows, time-dependent travel times, and Industry 4.0 data. The network is defined as a time-expanded directed graph, utilizing data from positioning systems, cargo sensors, traffic, and dynamic population distribution to update travel time and risk parameters. Four objective functions are defined: minimizing total cost, expected risk, CO2-equivalent emissions, and service delay/instability. To solve the problem, a problem-specific version of NSGA-III is introduced—incorporating route encoding, constraint handling, reference-point-based selection, and local search—and compared against NSGA-II, MOEA/D, and MOPSO. The primary innovation lies in integrating a FIFO-compliant time-dependent model, vehicle-cargo compatibility, dynamic population exposure, and a data-to-decision architecture. This paper does not present computational results or claims of empirical superiority; Rather, it provides the model, algorithm, reproducible experimental design, and statistical analysis method for future empirical testing.

Keywords : hazardous materials heterogeneous fleet routing industry 4 0 time window time dependent network nsga iii multi objective optimization