A Class Of Customer Relationship Management Optimization Model And Its Solution

The task comes from the project of CRM (Customer Relationship Management) One to One marketing optimization algorithm and optimization software development, which cooperate with the Tepsys Company in Japan. One to One marketing optimization is always attention. Because One to One marketing optimization problem is the large-scale programming problem, and the determination of the optimization model involves many marketing rules, it is impossible to consider all factors in one and the same optimization model as well as it is difficult to establish the One to One marketing optimization model that approach in fact and can be solved. In addition, the restriction of running time and memory consuming, most of optimization algorithms can't solve the large-scale One to One marketing optimization problems. For this reason, there is a necessity to study the One to One marketing optimization model and optimization methods.The thesis first analyzes a simple case of One to One marketing optimization problem to establish a set of One to One marketing optimization model. With the analysis of the optimization model, the objective functions and the constraint conditions are perfectly determined according with real-life marketing, and thus the new One to One marketing optimization model can meet the requirement of real-life marketing process. Secondly, with the new established optimization model, the optimization methods are studied in data treating, modeling method, algorithm optimization proposal and software realization. From the optimization methods, the One to One marketing optimization algorithm based on Minimum Degree Permutation is introduced in three parts: its basic idea, its narration and its realization.In order to evaluate the One to One marketing optimization algorithm based on Minimum Degree Permutation, a popular optimization sofeware package, GLPK, is tested with the software WHCRM developed by us. It concludes that the One to One marketingoptimization algorithm based on Minimum Degree Permutation is superior to GLPK. In addition, this thesis also describes the testing goal and testing method.