Group-based Adaptive Differential Evolution For Chance Constrained Portfolio Optimization Using Bank Deposit and Bank Loan

摘要

Portfolio optimization using bank deposit and loan is formulated as a chance constrained problem in which a non-risk asset called bank deposit is included in the portfolio and the borrowing money called bank loan can be invested in risk assets. The chance constrained problem is transformed into an equivalence problem. Then an Adaptive Differential Evolution using Directed mutation (ADED) is proposed and applied to the equivalence problem. Furthermore, in order to cope with large scale instances, ADED is extended to group-based ADED called ADED4G. ADED4G divides the feasible region of the chance constrained problem into four areas and executes ADED in each of the areas simultaneously. Experimental results show that both ADED and ADED4G are excellent. Besides, the practical use of bank deposit and bank loan improves the efficient frontier.