Computational Financing Techniques and Fundamental Challenges in Portfolio Optimization

Abstract

It is very fundamental to note that scientific computational financing is a rapid growing field, especially in the area of investment management. Considering the opinions of various researchers in this field, it is still very obvious that there are many unresolved challenges. These challenges had constituted limitations to the financial securities portfolio managers over the years in optimizing a particular objective function. It is on this foundation, that this paper reviewed relevance of computational financing in portfolio management, especially as it affects decisions on investment diversification, price option, risk management, investment viability and acceptability. The paper also identified some complex challenges of computational financing in optimizing portfolio such as capital allocation among alternative opportunities, portfolio constraints, taxes computation and investor’s portfolio preference. It was discovered that the conventional assumption of Markowitz does not hold which consequently leads to inefficient portfolio optimization performance. This paper, therefore suggested that the mean and variance of expected return should first be predicted with Functional Link Artificial Neural Network Model. The value of mean and variance generated will further serve as input in multi objective swarm intelligence techniques for better performance of portfolio optimization.