Fuzzy Clustering-based Transferability of QAOA Parameters for Solving Vertex Cover Problems
Giovanni Acampora, Angela Chiatto, Autilia Vitiello · 2025
Optimization problems are fundamental to real-world applications, driving the need for increasingly advanced computational techniques. One promising methodology is the Quantum Approximate Optimization Algorithm (QAOA), a hybrid quantum-classical approach that maximizes the probability of obtaining sub-optimal solutions for combinatorial problems by tuning quantum gate parameters with a classical optimizer. Despite its success, QAOA still faces practical challenges, particularly in terms of training efficiency and computational cost. This paper introduces a fuzzy clustering-based approach to transfer QAOA parameters from instances of a relatively simpler problem to instances of different, more complex problems. Experimental results demonstrate both the effectiveness and efficiency of this approach in solving the Vertex Cover problem, thereby offering a viable alternative to conventional parameter tuning methods.