Computational Game Theory

COMP34612 Computational Game Theory 10 Semester 2 Undergraduate Level 3 Xiaojun Zeng Unit coordinator Department of Computer Science 1 FHEQ level (Framework for Higher Education Qualifications) ' Last part of a Bachelors ' European Credit Transfer & Accumulation System Rating : 5.0 There has been a substantial grow of research activity at the boundaries of game theory, artificial intelligence, economics, computer science, and a number of other disciplines in recent years. The reasons behind this are twofold: On the one hand, game theory and its applications raise many important and challenging computing, learning, and communication problems to CS and AI; On the other hand, game theory provides important insights and powerful frameworks to a number of CS topics, including AI, Multi-agent systems, computer networks as well as many others. The main contents of this module include:1) To introduce the concepts and computational solutions for non-cooperative and cooperative game theory with their applications2) To introduce the machine learning techniques to solve the learning issues arise from the applications of game theory with their applications3) To introduce the mechanism design (the reverse game theory) and its applications for the design of the rules of a game The module includes a major piece of coursework (a group project run over 5 weeks) to apply game theory and learning methods covered to solve the pricing game problem.  There has been a substantial grow of research activity at the boundaries of game theory, artificial intelligence, economics, computer science, and a number of other disciplines in recent years. The reasons behind this are twofold: On the one hand, game theory and its applications raise many important and challenging computing, learning, and communication problems to CS and AI; On the other hand, game theory provides important insights and powerful frameworks to a number of CS topics, including AI, Multi-agent systems, computer networks as well as many others. The main contents of this module include:1) To introduce the concepts and computational solutions for non-cooperative and cooperative game theory with their applications2) To introduce the machine learning techniques to solve the learning issues arise from the applications of game theory with their applications3) To introduce the mechanism design (the reverse game theory) and its applications for the design of the rules of a game The module includes a major piece of coursework (a group project run over 5 weeks) to apply game theory and learning methods covered to solve the pricing game problem.  This module teaches the fundamental concepts of game theory and their computational methods to enable students to master the concepts/tools from game theory to model/analyse the interaction agents/systems, and to build skills in machine learning and optimisation methods for game analysis and problem solving. ILO 1: Cooperative game theory: Master the basic concepts and computational methods of cooperative game theory, including to be able to carry out the decision analysis to form coalitions, define transferable utility characteristic function to establish the relationship between a coalition and its value, choose the right properties/requirements (such as fairness, efficiency, marginal contribution), and apply and calculate the important Shapley value to make the decisions in fair surplus distribution, resource allocation, and cost share, when given a simple application problem.ILO 2: Learning and playing in games: Able to analyse the needs for and scenarios of learning with a given game with imperfect information or bounded rational behaviour. For a given simple application problem, able to formulate the reaction function learning problem, choose/apply right learning methods/algorithms, and then find the best strategy based on the learning outcome.ILO 3: Mechanism design: Master the concepts and computational methods of mechanism design. Able to analyse, categorise, and define a mechanism design problem, when given an application problem. Able to analyse and solve some simple application mechanism design problems and design the right pricing rules and application rules in order to achieve the best outcomes.ILO 4: Selected Applications of Game Theory: Able to assemble the above skills to solve some simple practical application problems such as competitive pricing games and carry out the analysis/evaluation of the outcomes.ILO 5: Non-cooperative game theory: Master the basic concepts and computational methods of non-cooperative game theory with perfect information, including Nash games and Stackelberg (leader-follower) games. For an application problem of Nash and Stackelberg games, able to establish the game framework, determine the game rule setting, identify the appropriate information structures, formulate payoff functions, determine strategy spaces, identify the right solutions based on the concepts of equilibrium/best strategy, and compute the best strategies by applying the derivative method. Workshop: 5 hoursTutorial: 6 hoursLecture: 12 hoursPractical: 30 hoursIndependent Study: 47 hours  1 Written exam 50% 2 Written assignment (inc essay) 50% COMP11120 Mathematical Techniques for Computer Science Pre-Requisite Compulsory COMP13212 Data Science Pre-Requisite Compulsory COMP24112 Machine Learning Pre-Requisite Optional COMP34111 AI and Games Co-Requisite Optional MATH11121 Mathematical Foundations & Analysis Pre-Requisite Compulsory MATH10111 Foundations of Pure Mathematics B Pre-Requisite Compulsory COMP34111 co-requisites (desired) COMP24112 (desired) MATH11121 or equivalent (for CM students), COMP11120 (for CS students), COMP13212 (all compulsory)COMP34111 co-requisites (desired) COMP24112 (desired)  MATH11121 or equivalent (for CM students), COMP11120 (for CS students), COMP13212 (all compulsory) N Lectures 12 Practical classes & workshops 35 Tutorials 6 0 47