Enhancing Stock Investment Returns with Learning Aggressiveness and Trust Metrics
on a daily basis , and the rate values come from the Federal Reserve Web site . In addition , the transaction cost is set to be $ 10 for each transaction . As a result , agents cannot capture the niche profit with high volume of buying and selling in slight price change situations . At the same time , transaction costs provide agents with a stable trading environment .
4.1 Agents
The world of stock trading in this ABM simulation is built from a collection
of agents . In order to reduce the complexity of the problem and to minimize the size of the exploration space of the model as much as possible , we decided to investigate agents on the individual level only . Although one of the major participants in the financial markets is a group known as institutional investors , our intent was to find the optimal balance between learning aggressiveness and trust metrics for individual agents only . The concept of individual investors refers to common investors in the stock market .
Agents are initialized with a pre-specified amount of money and with randomized initial trading strategies . Once they start trading , their available capital , trading strategies , and market momentum influence and trigger transactions . The momentum in the market adjusts agents ’ decision rules temporarily , and the individual trust metrics decides the degree of change . Agents also have the opportunity to learn the best trading strategies from the best performers within their radius during the open learning period . The speed of learning is controlled by their aggressiveness value . On the basis of Su and Hadzikadic ( 2014 , 2015 ), Table 1 describes the trading rules assigned to individual agents .
Table 1 . Trading Rules Assigned to Individual Agents
53