A Multidimensional Knapsack Model For Portfolio Optimization In Asset-backed Securitization

In recent years, financial leasing companies have developed fast. The special running method makes these companies need more fund, however, because the way of financing is solitary, which limits the development of these companies. Therefore, financial problem is the problem which needs to be solved first. Asset-backed securitization is one of the most important financing channels for financial leasing companies and the financial costs is the key factor that influences the corporate financing of these companies. Through properly selecting the basic assets and optimizing the assets pool, asset-backed securitization originator could reduce the cost of financing. Hence, the research problem is how to maximize the gains of originator by selecting basic assets.In this article, the related literature and theories are summarized. The research topics mainly focus on the development, motivation, supervision, venture, legislation and pricing of asset-backed securities. The research problem in this article was lack of study, only Mansini R et al analyzed the problem, they assumed respectively that the refund mode of basic assets is matching principal installments or matching service installments, which means the refund mode is single. This assumption does not match reality because the refund modes of basic assets are variety. Furthermore, they believe the foregone earnings of originator are the average value of foregone earnings at every stages, this kind of method is simple but does not satisfy the time value theory. Therefore, the problem requires further considerationBased on the present studies, the stakeholders profit are considered as constraints, the originator earnings as target and we introduce the time value theory to the basic assets, finally, we construct the multi-dimension knapsack problem Because it is hard to get the eligible data of financial leasing companies and our research problem is not the present situation of asset-backed securities, therefore, only simulation is carried out using Matlab R2011b to verify the validity of the model. Based on the characteristics of easing asset,200 easing assets are randomly generated and meet the related constraints. Simulation results show that the model is effective and the originator should synthetically considers all the parameters when choosing basic assets. The refund modes have little influence on asset value but the residual repayment stages influence the asset value. The closer residual repayment stages to ban repayment date, the lower foregone earnings and the higher profit of originator. The variety of basic assets can make sure every participator win-win, therefore improve the success rate of securitization and improve the profit of originator.