A Metaheuristic-based DSS for Portfolio Optimization

In this paper we present a framework, i.e. a concept and design as well as results with a prototypical implementation of a metaheuristic-based decision support system PM-DSS c © for portfolio optimization and managing investment guidelines. PM-DSS c © can be used for active as well as passive fund management. 1 The general problem In this paper we present a framework, i.e. a concept and design as well as results with a prototypical implementation of a metaheuristic-based decision support system PM-DSS c © for portfolio optimization and managing investment guidelines. PM-DSS c © is a DSS-generator, i.e. specific decision support systems for active as well as passive fund management can be customized using the models and methods which are packaged in PM-DSS c ©. For an introduction into modern portfolio theory see [3] and [4]. In [1] a concept of optimization in passive portfolio management is discussed. The decision situation that we assume is the following: After the initial design of a portfolio the fund manager is constantly confronted with the following decision problem of Portfolio Re-Optimization: Due to changes on the market the given portfolio has become infeasible because of violations to the investment guidelines and/or sub-optimal due to under-performance with respect to yield, risk etc. In this daily situation the fund manager has to restructure the portfolio through a series of sell and buy transactions. Here the transactions themself may be under some restrictions as for instance minimal/maximal amounts etc. There are highly sophisticated software systems on the market which routinely check portfolios as well as transactions for violations of constraints from investment guidelines. Yet, still the fund manager has no support in identifying promising transactions which will result in a feasible and profitable portfolio.