Mozilla Skin

Genius

From Negotiation

Enlarge

We have developed a negotiation environment that implements an open architecture for heterogeneous negotiating agents. It provides the basis for an implementation of a testbed for negotiating agents that includes a set of negotiation problems for benchmarking agents, a library of negotiation strategies, and analytical tools to evaluate an agent's performance and their strategies allows user. You can download the latest version, Genius 3.2.1. This includes the user guide explaining how to use it, and a number of negotiation domains and negotiating agents. You can use the source code from an example agent as a starting point.

You can watch a movie about how to use the software here.

The software requires a Java Runtime Environment, version 1.5 or newer. It can be run on the most of the platforms that support Java 1.5.

The software was demonstrated at the Twelfth International Workshop CIA 2008 on Cooperative Information Agents (CIA'2008) that was held in Prague, Czech Republic, where it won the "System Innovation Award".



Research

Genius is actively used in academic research. If you want to refer to Genius in your paper please use the following paper:

Genius: An Integrated Environment for Supporting the Design of Generic Automated Negotiators (pdf) (BibTeX).


For examples of uses of Genius in research, see:

  • The first automated negotiating agents competition (ANAC 2010).(pdf) (bib)
  • Genius: An Integrated Environment for Supporting the Design of Generic Automated Negotiators (pdf) (BibTeX).
  • Supporting the Design of General Automated Negotiators (pdf)

Releases

Genius 3.2.1 (download)

Used in ANAC 2012, The Third International Automated Negotiating Agent Competition.

For ANAC 2012, we have extended Genius with a classic feature in negotiation that was absent in previous years, namely reservation values.

A reservation value is a real-valued value that sets a threshold below which a rational agent should not accept any offers. Intuitively, a reservation value is the utility associated with a Best Alternative to a Negotiated Agreement (BATNA). A reservation value is the utility that an agent will obtain if no agreement is realized in a negotiation session. This can happen either by breaking off the negotiation (a non-zero reservation value makes breaking off potentially interesting) or by not reaching an agreement before the deadline associated with a negotiation session (if such a deadline is present). Reservation values typically are different for each of the negotiating agents that negotiate in a session. In case no reservation value is set in a profile, the reservation value of an agent is set to zero (0). Given that utilities for outcomes range from 0 to 1, this essentially is the same as having no reservation value. If a discount factor is present, reservation values will be discounted in exactly the same way as the utility of any other outcome.


Genius 3.2 (download)

Changelog:

+ All ANAC 2011 resources are included

All agents and negotiation scenarios of ANAC 2011 are included in the repository of this release.

+ User Interface Improvements

Many improvements have been made to the GUI, including highlighting of the last bid, an improved depiction of the outcome space, and a more user-friendly selection panel of the scenarios, agents, and protocols.

+ Previous tournament setup is remembered

The tournament setup is automatically saved every time you run a tournament.

+ Multiple sessions

It is now possible to run multiple instances of the same negotiation session.

+ Better logging

It is possible to save negotiation logs to .xml format. Negotiation features that can be logged include: time of agreement, number of exchanged bids, negotiation domain and preferences, opponent information, and distance of agreement to Nash and Pareto.



Genius 3.1 (download)

Used in ANAC 2011, The Second International Automated Negotiating Agent Competition.

Changelog:

+ Changed protocol so that agents now have a shared time line of 3 minutes

The real-time deadline of ANAC is considered to be a challenge for the competitors. In 2010, the agents had three minutes each to deliberate. This means agents have to keep track of both their own time and the time the opponent has left. For ANAC 2011, we have chosen a simpler protocol where both agents have a shared time window of three minutes.


+ Discount factors

ANAC 2011 will have domains that have discount factors. In ANAC 2010, almost every negotiation between the agents took the entire negotiation time of three minutes each to reach an agreement. Adding discount factors should provide more interesting negotiations with faster deals. The implementation of discount factors is as follows. We have added a method to UtilitySpace that gives the discounted utility:

public double getUtilityWithDiscount(Bid bid, Timeline timeline)

and a convenience method in Agent that does the same thing:

public double getUtility(Bid bid)

Its definition is as follows. Let d in (0, 1) be the discount factor. (If d<= 0 or d>= 1, we assume that d = 1.) Let t in [0, 1] be the current normalized time, as defined by the Timeline. We compute the discounted utility discountedUtility as follows:

discountedUtility = originalUtility * d^t.

Note that for t = 0, or d = 1, the utility remains unchanged.

When we use normalized time, then for t = 1 the original utility is multiplied by the discount factor. The effect is almost linear in between.



Genius 3.0 (download)

Used in ANAC 2010, The First International Automated Negotiating Agent Competition.

The first public version of Genius. Uses a separate timeline for the negotiating agents.



Genius 4.0 (unreleased)

Please be aware that a major new release of Genius is currently being developed. The framework of Genius 4.0 has been completely revised, which may cause compatibility issues with current additions to the Genius code and with custom resources such as agents, domains and preference profiles. For more information, please contact us.

Automated Negotiating Agents Tournament

Genius is used in the annual Automated Negotiating Agents Competition (ANAC)


Previous Competitions