Credit score card threshold optimization strategy based on simulated annealing algorithm

When banks use credit score cards, they need to find the best combination among multiple credit score cards and thresholds to maximize the final revenue. In this study, we propose a credit score card threshold optimization method based on simulated annealing algorithm to address this problem. First, we transform the problem into QUBO (Quadratic Unconstrained Binary Optimization) form and introduce Lagrange multipliers to handle the constraints. Then, a simulated annealing algorithm is used to find the optimal credit scorecard and threshold combination. In addition, we reduce the computational time complexity by an approximation algorithm. This study provides an efficient credit scorecard threshold optimization method for banks, which helps to increase their final revenue and reduce bad debt losses.