Lecture notes will be available individually with sources on piazza A book of lecture notes pdf is available here.

===========================================================

T7/R78; office hours Tues 8th period or by email appointment

Prerequisites:
===============
The course is meant for upper division and graduate students who have had **ANY TWO** of the following types of courses:
--discrete mathematics (or combinatorics)
--data structures and algorithms;
--mathematical programming and optimization (or operations research);
--game theory OR microeconomics foundations
--complex/dynamical systems.

There is no single text. The one that comes closest is available online free at (click on the resources tab and then the text under "general resources" after that one more click to get the pdf): Nisan-Roughgarden-Tardos-Vazirani>

Course Description
====================
The course is motivated by various internet and peer-to-peer economies. We will use formal, rigorous modeling, with proofs throughout. There will be significant emphasis on computational, specifically algorithmic and complexity issues -- familiarity will be assumed.

The course will train students to algorithmically model and analyze "microeconomic games" consisting of a permitted set of moves/interactions/transactions by and between "players." Each player has an individual utility or satisfaction function whose maximization directs a player's (local) choices and behavior. However, the utility function depends not only on the player's own choices directly, but also globally on all other player's choices, which are in turn motivated by maximization of their own utilities. We will consider a variety of examples of game set up and rules and utility functions, with economic, philosophical or even evolutionary biological relevance.

We will look at (as time permits) standard marketplaces, auctions (Vickrey, google adwords pay-per-click, ebay), voting, distributed/network QoS/resource allocation settings; credit and trust networks (ripple); reputation economies (facebook likes, etc.); barter/reciprocity economies (couchsurfing), opensource economies (that generate a ton of content) as well as peer-to-peer filesharing and other economies.

The game set up could be on a "peer to peer network" or on a "broadcast network/open marketplace;" the game could involve strategy or be strategy-proof. Utility functions, while individual, need not be purely "selfish:" they can encode an individual's wish for the collective or other types of fairness affinity preferences or altruism.

We will be interested in the characteristics of various types of equilibrium/optimum states of the game (e.g. Nash/Pareto), how they depend on the rules of the game and the utility functions, the computational complexity of the game for reaching equilibrium/optimum states, as well as the effect of the player's computational complexity; we will also study the game as a simple dynamical system, including phenomena such as phase-transitions, percolation and chaotic behavior as well as instability/failure.

We will study various "problems" in games despite all players following rules (i.e., bad equilibria, "undesirable" but legal behaviors by players, e.g. freeloading) and how to use rigorous analysis to design tweaks to the game rules to prevent such problems (mechanism design). Finally, as time permits, we will attempt to formalize classical adages such as: "boom and bust cycles are unavoidable," "growth results in inequality" "inequality is required for the economic engine to function" "greed is necessary for economic activity" etc. and their relevance in internet based economies (that assume a substantial availability of information, communication and computational power). As time permits, we will seek rigorous, formal answers to whether the internet age ameliorate or exacerbate these adages?

Material
==========

Not all of the material we will cover is available as books or lecture notes. That said, almost all the material is on the web.

The course will be held together by a thread of classical (micro) economics, from
Kakutani's fixed point theorem
to convexity-based works by Nash and Walras and (Nash/Walras version directly using Brouwer's fixed point theorem)
and Arrow-Debreu, see also (here)
and alternatives like Rabin's notion of fairness,
work by Fehr-Schmidt on cooperation and fairness
and on free-riders
also in the p2p context
network p2p economies p2p economics in general
and on reputation

More recently, there are books on Algorithmic game theory, network economies and related topics,
http://www.cs.cornell.edu/home/kleinber/networks-book/
http://www.cambridge.org/us/academic/subjects/computer-science/algorithmics-complexity-computer-algebra-and-computational-g/algorithmic-game-theory?format=HB >

such as computational complexity of computing fixed points and equilibria
http://theory.stanford.edu/~tim/papers/et.pdf>
http://www.cs.berkeley.edu/~christos/equi.ps>

and several lecture notes. Here are some examples, but there are plenty more available on the web. recent videos of Roughgarden's course
a primer
http://www.eecs.harvard.edu/~parkes/mechdesign.html
http://users.eecs.northwestern.edu/~hartline/amd.pdf
http://people.csail.mit.edu/costis/6853fa2011/
http://p2pecon.berkeley.edu/

Grading and Expectations
======
(4930 and 6930 students will have different bars while grading)
--2 tests
--class participation and preparing precise/web-publishable lecture notes each student for a 1-week segment (prepare to make many edits)
--1 final project/paper/presentation
After the first 4 weeks people should propose topics (and teams) for papers and projects and start thinking about them.

NOTE: While we will try to give background and definitions as time permits, students should be prepared to acquire the required background on their own. Talking to your classmates or the instructor will give you good sources for the background.

Prepare to work on at least 6-7 medium-hard to hard problems per week from the assigned homeworks or problems from NRTV book and to spend 9 hrs outside class studying. This will not be graded but will be required for the tests. Additional time may be needed for the lecture note preparation and projects.

Please contact the instructor if you have questions: sitharam@cise.ufl.edu

Emphases/Style of Course
===================
The emphasis of the course will be meaningful formal microeconomic models with algorithmic content and rigorously provable results. We will spend (not necessarily in this chronological order) atleast 1/3 of the time on concepts and theorems in classical microeconomics with standard examples of economies. And at least 1/2 time on an "algorithmic" version/analysis of each question, about equilibria, existence, uniqueness, optimality, mechanisms towards the above as well as towards efficiency (rapidity of convergence) both with classical examples of economies as well as other standard ones in the internet age (auctions, networks). And then if time permits, some time on questions (fairness, reputation, free-riding) where a computational/information/communication resource should be introduced into the set-up and analysis (but has not been done). Any further available time will be spent on various less-studied economies, either emerging through the internet (credit and trust(ripple), barter(couchsurfing). bitcoin etc.), or simply computationally less-studied economies (such as evolution).

Lecture schedule (about 2-4 weeks each)`
======================

(1) Chapter 1 and 2 and 9 and 17 (overview chapters) of the NTRV book with many examples of games and economies; illustrations of different settings classical and now standard for the internet, examples and mathematical types of games; different types of equilibria; recurring issues of existence, optimality, efficiency/complexity; and recurring concepts such as mechanism design towards existence, optimality and efficiency. This will be accompanied with detours to relevant classical microeconomic papers and results and some of the background for the theorems that they use.

(2) Complexity of Equilibria (material from several subsequent NRTV chapters) (question of efficiency of computation of equilibria); knowledge of complexity classes P and NP and some familiarity with the notion of "reduction" of one problem to another will be important here.

(3) Mechanism Design (second and 3rd sections of NRTV) - towards existence, optimality, efficiency/complexity of equilibria

If time permits: (4) Fairness/Cooperation/Reputation/Punishment for freeriding investigation of computational issues here

If time permits: (5) nonstandard economies - modern internet-based; or unusual settings/scenarios modeled as economies

Instructor Sitharam's papers on this general topic
========================
A. Lomonosov, M. Sitharam ``Stability, Optimality and Complexity in Network games with pricing and player dropouts,'' SCPE (Scalable computing) journal, 2007
A. Lomonosov, M. Sitharam, K. Park ``Stability in network games'' Journal of Computer and System Sciences 2004.
Kihong Park, Meera Sitharam, Shaogang Chen, Quality of service provision in noncooperative networks with diverse user requirements, (pdf file) Decision Support Systems, Special Issue on Information and Computation Economies, 2000.
Shaogang Chen, Kihong Park, Meera Sitharam, On the Ordering Properties of GPS Routers for Multi-Class QoS Provision, (pdf file) SPIE Conference on Performance and Control of Network Systems, Nov. 1998. Kihong Park, Meera Sitharam, Shaogang Chen, Quality of Service Provision in Noncooperative Networks: Heterogenous Pref erences, Multi-Dimensional QoS Vectors, and Burstiness, (pdf file) Proceedings of the International Conference on Information and Computation Economies, ICE 1998.