The development of constrained optimization algorithms for route planning in road transportation requires an innovative approach that must be flexible (e.g., capable of optimizing different utility functions) and adaptable to dynamic and not totally predictable configurations of the environment (e.g., traffic congestion). .
In this context, the aim of the activities will be a comparative analysis and the design of innovative solutions based on evolutionary algorithms for the generalized and multi-objective Vehicle Routing Problem (VRP). The analysis will consider how the main features of the problem (e.g., graph topology, selected performance metrics etc.) influence the solutions that the algorithms provide considering both the quality of the solution and the computational aspects (e.g., time and memory needed to execute the algorithm).
The development will mainly focus on innovative methods for the parameterization of state-of-the-art approaches through the use of solutions based on machine learning.