Sharpe Index Based Portfolio Optimization Using Computational Intelligence

Portfolio optimization is a complex problem with multiple objectives, namely maximizing investment returns while simultaneously minimizing variance (risk). Particularly when considering real-time constraints, it becomes a nonlinear problem that can be effectively addressed through the utilization of computational intelligence techniques. This research paper introduces two such techniques, namely Particle Swarm Optimization (PSO) and Genetic Algorithms (GA), for optimizing the portfolio of the Karachi Stock Exchange 30 index. The selection model for the portfolio is based on Markowitz's mean-variance theory, enhanced with floor and ceiling constraints, and considers expected returns and the Sharpe ratio. The results obtained from these computational intelligence techniques are compared with the optimization capabilities of MS Excel (Solver). The findings reveal that PSO outperforms both Solver and GA in terms of achieving superior optimization outcomes.