Chapter 8 Introduction To Multiagent Systems

Front 1

CDPS

Back 1

Cooperative Distributed Problem Solving

Front 2

benevolence assumption

Back 2

The assumption that agents will always be helpful, will always tell the truth: in general, where agents have conflicting preferences or goals, we cannot make the benevolence assumption.

Front 3

parallel problem solving

Back 3

The idea of exploiting parallelism in problem solving.

Front 4

task sharing

Back 4

An approach to cooperative problem solving in which an overall task is decomposed into smaller tasks, which the 'task manager' then advertises to agents, who can then bid to carry out the tasks.

Front 5

result sharing

Back 5

A form of cooperative problem solving, in which a partial solution to problems are shared by agents.

Front 6

Contract Net

Back 6

CNET - A classic protocol for task sharing, modelled on the process of putting contracts out for tender in human businesses.

Front 7

task announcement

Back 7

In the contract net protocol, a task announcement is sent out by a prospective manager to potential bidders, defining the task that is to be carries out.

Front 8

directed contract

Back 8

A variant of the contract net pool.

Front 9

task announcement processing

Back 9

na

Front 10

bid processing

Back 10

One of the key stages of the contract net protocol: upon receipt of bids for a task by multiple agents, the process of choosing between them.

Front 11

award processing

Back 11

na

Front 12

award and inform processing

Back 12

A key stage in the contract net task sharing protocol: handling the awarding of tasks to agents, and informing agents of their success or otherwise in bidding for tasks.

Front 13

marginal cost of task

Back 13

The cost involved in taking on a new task.

Front 14

FA/C

Back 14

Functionally Accurate / Cooperative

Front 15

coordination

Back 15

The problem of managing the interactions between the actions of multiple agents.

Front 16

coordination relationships

Back 16

In von Martial's typology of interactions between actions, coordination relationships capture ideas such as one action enabling or disabling another

Front 17

action equality

Back 17

na

Front 18

consequence

Back 18

na

Front 19

favour

Back 19

A type of coordination relationship: an action favours another if it enables it, or improves it in some way.

Front 20

DVMT

Back 20

Distributed Vehicle Monitoring Testbed - an early multi-agent distributed sensing system.

Front 21

partial global planning

Back 21

A coordination regime for cooperative problem solving

Front 22

GPGP

Back 22

Generalized Partial Global Planning - an approach to coordination in multiagent systems.

Front 23

commitment

Back 23

a.k.a. intentions: something the agent has resolved to achieve.

Front 24

convention

Back 24

na

Front 25

motivation

Back 25

na

Front 26

acquaintance models

Back 26

A data structure capturing what an agent knows about its peers.

Front 27

social laws

Back 27

A social law is a set of constraints on the behaviour of agents, designed so that if these constraints are adhered to, then some overall desirable objective results.

Front 28

strategy update function

Back 28

A function that tells an agent how to modify its strategy on the basis of what it has whitnessed so far.

Front 29

useful social law

Back 29

na

Front 30

focal state

Back 30

In social laws, a focal state is a state that an agent wants to be able to visit without being prevented from doing so.

Front 31

useful social law problem

Back 31

the problem of checking whether there exists a social law that, if implemented, would enable all agents to achieve their goals - identified and proeved NP-complete by Shoham and Tennenholtz.