Research on the optimization of credit score card for solving the QUBO model based on annealing algorithm and quantum computer

The optimization algorithm of the quantum computer can better find the correlation between the different factors in the credit score card, so as to better predict the credit risk. By establishing a multivariate nonlinear integer programming model and transforming it into a QUBO model [1], the bank credit score card is combined optimized to maximize the total bank revenue. First, we establish a multivariate nonlinear integer programming model to convert the quadratic Hamiltonian operator into the form of multiple matrix products. We use the quantum annealing algorithm to find a credit score card in 100 credit score cards and find the corresponding threshold to maximize the total revenue target function. Secondly, on the basis of the above model, we choose the designated three credit score cards and only select one set of thresholds for each credit score card, and find the threshold values corresponding to the three credit score cards, so that the total income target function is the largest. Finally, we arbitrarily selected three credit score cards and only each credit score card chooses one set of threshold, and found the threshold corresponding to the three credit score cards, so that the total income target function is the largest. The above are converted to QUBO form solution [2].