Utilizing Ant Colony Optimization And Technical Analysis To Establish An Investment Model Of Taiwan's Stock Market

The Ant Colony Optimization (ACO) is proposed by Marco Dorigo in 1992. Then, ACO is applied to many industrial problems, such as Quadratic Assignment Problem, Network Routing Problem, Traveling Salesman Problem, and so on.More and more people invest or manage their money. The stock market is the most developed among many Taiwan's financial investment tools. Also the stock market is preponderant in safety, liquidity, and flexibility.There are many ways to predict the trend of a certain stock. The common two ways are fundamentals analysis and the technical analysis. The fundamental analysis is to analyze the essence of a certain enterprise; and the technical analysis is using some indexes to predict the stock price. However, the investment environment is variable. So how to select an investment combination becomes a very important topic for study.This study combines the ACO and the stocks'transaction volumes and investment returns. Base on the stock prices, there are three situations- risk avoider, risk taker, and risk neutral. The option to invest on the bank is to be as the hedge tool. If the investment returns are not good, the model we built will put the money into the bank. In this model, there are four investment targets: the first three are stocks from Taiwan's stock market which are Powerchip Semiconductor Corporation (PSC), Formosa Petrochemical Corporation (FPCC), and Chinatrust Financial Holding Company (CFHC); the last one is bank savings for hedge purpose.This study puts the real stock data from Taiwan's stock market from January 2004 to April 2005 into the model, and the results are shown as the following:Firstly, the return on investment of the model is 17.5% better than the standard system.Secondly, the decision parameters of risk avoider, risk taker, and risk neutral we proposed in the model would have certain impacts on the investment returns.Thirdly, the differentα,βwould have impacts on the investment returns.Fourthly, after the model's simulation, it had shown a better result using the investment targets we set in the model.