The PhD research investigates algorithmic frameworks for solving NP-hard routing and scheduling problems formulated as constrained combinatorial optimization problems. The problem classes considered are defined over discrete solution spaces with complex feasibility regions, multiple objective functions, and time-dependent constraints, as commonly encountered in transportation, traffic, and mobility systems.
The work develops metaheuristic and hybrid optimization algorithms, including but not limited to evolutionary algorithms, simulated annealing, tabu search, variable neighborhood search, and memetic algorithms that integrate exact local search procedures. Problem instances are modeled using graph-based and permutation-based representations, while solution quality is evaluated through single- and multi-objective fitness functions. Constraint handling is addressed using penalty functions, repair operators, and feasibility-preserving search operators. The research also considers online and dynamic optimization settings.
A major contribution lies in adaptive algorithm control, where algorithm parameters are treated as dynamic variables rather than static constants. Since transportation and mobility problems exhibit non-stationary behavior, due to fluctuating demand, changing network conditions, and stochastic disturbances, self-adaptive and feedback-driven parameter control strategies are proposed. These strategies adjust parameters such as mutation and crossover rates, neighborhood size, tabu tenure, and cooling schedules based on convergence indicators, diversity measures, and fitness improvement rates. This enables algorithms to track moving optima and maintain search efficiency in dynamic environments.
Algorithm performance is analyzed experimentally using benchmark problem sets and realistic traffic and mobility scenarios. Comparative evaluation focuses on convergence speed, solution optimality, robustness, scalability, and computational complexity. Parallel implementations and hardware-oriented acceleration techniques are also investigated to enable near real-time optimization in large-scale systems.
The research advances the state of the art in adaptive metaheuristic optimization by providing algorithmic mechanisms for dynamic parameter control and by demonstrating their effectiveness in solving complex, dynamic routing and scheduling problems relevant to modern traffic and mobility applications.