基于QUBO模型的信用卡最优获利组合规划  

Optimal profit combination planning for credit cards based on QUBO model

作　　者：王文鹤 杜汉铭 WANG Wenhe;DU Hanming(School of Mechatronic Engineering,Changchun University of Technology,Changchun 130012,China)

机构地区：[1]长春工业大学数学与统计学院,吉林长春130012

出　　处：《长春工业大学学报》2024年第4期362-369,共8页Journal of Changchun University of Technology

基　　金：吉林省发改委基本建设基金项目(2022C043-2)。

摘　　要：在银行信用卡或相关的贷款业务中,为最大化银行收益、方便银行对客户授信,针对不同信用卡设定不同阈值的情况下进行投资组合,对模拟数据进行归一化处理,建立信用卡最优投资组合二次无约束二值优化(QUBO)模型,使用模拟退火算法求解模型得到单一信用卡与三张信用卡的最大收益阈值组合。进一步通过量子降维方法优化二次无约束二值优化(QUBO)模型,最终得到信用卡选择与阈值的双重最优组合,得出最大收益,辅助提高银行的业务竞争能力。In the business of bank credit card or related loans,in order to maximize the bank's income and facilitate the bank's credit to customers,we set different thresholds for different credit cards to conduct portfolio,normalize the simulation data,establish the Quadratic Unconstrained Binary Optimization(QUBO)model of the optimal portfolio of credit cards,and use Simulated annealing algorithm to solve the model to obtain the maximum income threshold combination of a single credit card and three credit cards.Further optimize the Quadratic Unconstrained Binary Optimization(QUBO)model through quantum dimensionality reduction methods,ultimately obtaining a dual optimal combination of credit card selection and threshold,obtaining maximum returns and assisting in improving the bank's business competitiveness.

关 键 词：二次无约束二值优化模型 模拟退火算法 组合优化 全局搜索 量子降维 

分 类 号：O212[理学—概率论与数理统计] O211[理学—数学]