Exploring Organizational-Learning Oriented Classifier System in Real-World Problems

Abstract

Learning Classifier Systems (LCSs) [12, 13, 14] are often compared with Reinforcement Learning (RL) [21]. Such comparisons suggest that many theoretical analyses have been studied in the context of RL, while few in the context of LCSs. However, LCSs are applied to many real-world problems, while RL is rarely applied. Examples of LCSs include aircraft maneuvers [10, 19], the controller or action planning of a physical robot [7, 20], trading in the stock market [17], electric power distribution networks [26], data mining from a large clinical database [11], the Wisconsin breast cancer dataset [29], and others [6]. These examples show the great advantage of LCSs in comparison with RL.