Portfolio optimization based on the behavior and risk appetite of the heterogeneous investor community in financial markets has been very difficult to model and predict accurately. In this paper, firstly we attempt to simulate a multi-agent based stock market; where different types of agents are modeled to trade stocks using various strategies. The observations from trading activity of the user are in turn used to assess the risk adversity level (RAL) by using a suitable fuzzy logic model. RAL score from the fuzzy model serves as input to perform portfolio optimization using Genetic algorithm. We further analyze and evaluate the optimum portfolio performance for different risk adversity level.
展开▼
