Course title
1M5180001
Agent Systems

 igarashi harukazu Click to show questionnaire result at 2018
Course content
Control of the entire system is not easy when attempting to accomplish multiple objectives in a large-scale system while considering complex constraints. For this reason, a distributed control method has been devised to control the entire system by describing the system through agent groups which act autonomously and by controlling the activities of each agent. In this course, students will learn about the basics and application of such “multi-agent systems.” The aforementioned themes will be studied through textbooks and exercises.
Purpose of class
To understand basic theories and skills on multi-agent systems and how to apply them in specific cases.
Goals and objectives
  1. Understand basic methods to solve problems on multi-agent systems.
  2. Solve simple examples using aforementioned methods.
  3. Acquire presentation abilities to explain aforementioned methods.
Language
Japanese
Class schedule

 Class schedule HW assignments (Including preparation and review of the class.) Amount of Time Required
1． Multi-agent systems Read the syllabus and read Chapter 1 of the text. 100minutes
2． Game theory (I): multi-agent systems and game theory; infinitely repeated game Read pp.16-26 of the text. 100minutes
3． Game theory (II): iterated prisoner's dilemma games; Nash equilibrium Read pp.27-39 of the text. 100minutes
4． Game theory (III): folk theorem; finitely iterated games Read pp.40-45 of the text. 100minutes
5． Agent learning (I): neural network learning Read pp.53-69 of the text. 100minutes
6． Agent learning (II): reinforcement learning framework Read pp.70-81 of the text. 100minutes
7． Agent learning (III): TD learning and Q learning Read pp.82-90 of the text. 100minutes
8． Evolutionary calculation (I): genetic algorithm Read pp.91-109 of the text. 100minutes
9． Evolutionary calculation (II): agent design example; schema theorem Read pp.110-122 of the text. 100minutes
10． Evolutionary calculation (III): genetic programming; classifier system Read pp.123-135 of the text. 100minutes
11． Soccer agent group coordinative behavior acquisition (I): agent formulation Read pp.136-148 of the text. 100minutes
12． Soccer agent group coordinated behavior acquisition (II): GA application and experiment results Read pp.149-155 of the text. 100minutes
13． Cases of multi-agent learning (I): pursuit problem Review the contents of the 6th and the 7th lectures. 100minutes
14． Q&A and vision for future multi-agent learning Review the contents of all the lectures. 1350minutes
Total. - - 2650minutes
Relationship between 'Goals and Objectives' and 'Course Outcomes'

 Report Total.
1. 33% 33%
2. 33% 33%
3. 34% 34%
Total. 100% -
Evaluation method and criteria
Report 100%.

Total score 60% is required to pass this course.
Textbooks and reference materials
Textbook:H.Ouchi, M.Yamamoto, and H.Kawamura, "Basics and Applications of Multi-agent systems", Corona-sha, 2002 (in Japanese).
Prerequisites
Basic knowledge of mathematics such as calculus and linear algebras.
Office hours and How to contact professors for questions
  • One hour after lectures.
Regionally-oriented
Non-regionally-oriented course
Development of social and professional independence
  • Course that cultivates a basic problem-solving skills
  • Course that cultivates an ability for utilizing knowledge
Active-learning course
More than one class is interactive
Course by professor with work experience
Work experience Work experience and relevance to the course content if applicable
N/A N/A
Education related SDGs:the Sustainable Development Goals
  • 9.INDUSTRY, INNOVATION AND INFRASTRUCTURE
  • 12.RESPONSIBLE CONSUMPTION & PRODUCTION
Last modified : Sun Mar 21 15:09:25 JST 2021