Learning Agents for Human Complex Systems
Author:
- Iris LorscheidILIris Lorscheid
- 1 July 2014
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 432-437
Abstract
Learning agents are a useful concept for agent-based simulation. The learning ability increases the autonomy of agents, which may lead them to unforeseen results on the individual (micro) and group (macro) level. Learning is even required to be successful in human complex systems, for which an adaptation to environmental changes is necessary. However, learning agents are considered as complex and not easy to understand. Therefore, they are not often applied in agent-based simulation models in the social sciences. This paper provides an overview of the learning agent concept by (1) putting learning agents in the context of the research field machine learning, (2) clarifying the basic learning agent decision process, and (3) providing a systematic overview of the learning agent properties as model guideline and communication scheme. This paper should encourage researchers in the field to apply learning agents by supporting the understanding and communication of the concept.Keywords
This publication has 24 references indexed in Scilit:
- Simulating Normative AgentsInternational Journal of Agent Technologies and Systems, 2010
- Checkers Is SolvedScience, 2007
- Using Decision Trees for Agent Modeling: Improving Prediction PerformanceUser Modeling and User-Adapted Interaction, 1998
- Learning Through Reinforcement and Replicator DynamicsJournal of Economic Theory, 1997
- Computer Simulation of Human ThinkingScience, 1961
- The simulation of verbal learning behavior1961
- Some Studies in Machine Learning Using the Game of CheckersIBM Journal of Research and Development, 1959
- An Iterative Method of Solving a GameAnnals of Mathematics, 1951
- I.—COMPUTING MACHINERY AND INTELLIGENCEMind, 1950
- A logical calculus of the ideas immanent in nervous activityBulletin of Mathematical Biology, 1943