=Paper=
{{Paper
|id=Vol-1876/paper03
|storemode=property
|title=A Proposal for Consensual Decision Making using Argumentation
|pdfUrl=https://ceur-ws.org/Vol-1876/paper03.pdf
|volume=Vol-1876
|authors=Ayslan Trevizan Possebom
|dblpUrl=https://dblp.org/rec/conf/ijcai/Possebom16
}}
==A Proposal for Consensual Decision Making using Argumentation==
https://ceur-ws.org/Vol-1876/paper03.pdf
Proceedings of CMNA 2016 - Floris Bex, Floriana Grasso, Nancy Green (eds)
A Proposal for Consensual Decision Making using Argumentation
Ayslan Trevizan Possebom
Federal University of Technology – Paraná, UTFPR, Curitiba-PR
possebom@gmail.com
Abstract the expertise of the issuer about the subject in discussion can
indicate the strength of the argument
We propose an approach where a consensual deci- This paper aims to propose an approach for decision mak-
sion making in multiagent systems can be reached ing using argumentation applied to multiagent systems where
using argumentation, with dialogues inspired in the agents dialogue with others by presenting their arguments in
interaction among humans in presencial meetings. favor or against about every decision alternative. The pro-
The goal of the study is to identify the main features posal contains three stages: (1) dialogue stage: a specific di-
needed to reach consensus, like speech acts that are alogue protocol is proposed to rule the sequence of moves
necessary in a dialogue using argumentation and and the speech acts. During the dialogue, all agents have a
ways in which agents can accept or reject formulas trust function that determines whether an agent must update
of the presented arguments and update their belief or not his knowledge base; (2) argumentation stage: a mech-
bases with those formulas. Every argument needs anism is proposed to calculate the strength of the arguments
a strength based on the acceptability of each for- based on the trust degree of the agents and the relation of ac-
mula. Our next step is to calculate the best decision ceptance/rejection of each formula in the argument; and (3)
among all the alternatives using the dialogues and decision stage: the definition of a strategy to calculate the
the strength of the arguments, and be able to com- preference relation about each decision alternative based on
pare our results to other proposals that do not focus the strength of the arguments and the attack relations.
on consensus. This work is organized as follows: Section 2 presents
the dialogue stage that has a framework for dialogue among
agents looking for consensus in each formula of the argument,
1 Introduction rules for dialogue moves, and the trust function used to up-
In presencial meetings, where there are a number of people dating beliefs with consensual information. Section 3 shows
discussing about an issue, the group needs to reach an agree- the need to calculate the strength of the arguments after the
ment on a single alternative among other available options. dialogue in the argumentation stage. Finally, Section 4 con-
By means of argument exchange, each individual dialogue cludes the work.
participant presents his opinion or points of view about the
issue under discussion. The decision making process consists 2 Dialogue Stage
of evaluating every information presented by the group dur-
Let A be a set of agents {a1 ,...,an } with n>1 that take part
ing the dialogue, endorsing or rejecting the entire arguments
in a dialogue, every agent ai uses a common propositional
or piece of the information in the arguments.
language L to represent their knowledge about the world. To
The most popular argumentation frameworks applied in
deal with dialogue among agents, we propose a protocol that
decision making are usually abstract (traditional, bipolar,
consists of:
question-and-answare, value-based) [Carstens et al., 2015],
others use an argumentation system with logical language • Arguments [Amgoud et al., 2002], and attack relations
[Muller and Hunter, 2012; Toni, 2014; Amgoud and Prade, of two types: undercut and rebutal [Parsons and McBur-
2009; García and Simari, 2004], or use social argumentation ney, 2003];
to represent votes in arguments or attack relations [Eğilmez • Structure of agents with knowledge base ⌃=Ki [ Gi [
et al., 2013]. These frameworks are about argumentation and KOji (Ki = knowledge about the environment, Gi = de-
decision making, but every one has a different mechanism
to achieve the best result. Neither of them deals specifically sirable features in a decision, and KOji = with i 6= j con-
with situations where consensus is required. To have con- sensual information from agent j) and a set of dialogues
sensus, we need to consider each formula (information) in an tables, one table for each decision alternative;
argument with the expertise (trust) of each issuer. The set of • Framework for consensual decision making that con-
formulas with their relations of acceptance or rejection and tains all elements necessary to execute the argumenta-
14
Proceedings of CMNA 2016 - Floris Bex, Floriana Grasso, Nancy Green (eds)
tion (agents, trust score of all agents, decision alterna- 4 Conclusions
tives, attributes considered with higher impact in the de- In this work we presented the current research of an ongoing
cision, waiting time, and a pre-order of the decision al- PhD thesis. The main goal is to investigate how a consensus
ternatives representing the result of the framework; can be reached among agents when they have to make deci-
• Artifact composed by a set of operations and set of ob- sions using argumentation, where all arguments in a dialogue
servable properties [Ricci et al., 2009]. This artifact con- have to be considered. We may detect to what extent each
tains a list that represents the queue with the sequence formula in the argument is believed by the group (consensual
of moves. An agent can: (1) register in the list that rep- information) and this relation to the strength of the argument.
resents the speech sequence; (2) query the list; and (3) We observed that when a formula of an argument is not re-
remove its registry in the list; jected, it may be reviewed by the agents and considered as
• Speech acts (Propose(Arg, Attack), Accept( , ), acceptable, leading to a consensual information.
Refuse( , ), Challenge( , Ag, h), Ask-if(Ag, ), We will continue with the analysis of the consensus on ev-
Query-if(Ag, µ, ), Inform( , Ag, )); ery formula and improve the calculus of strength regarding
the trust level of the agents and the relations of acceptances
• Trust function to analyse each formula in an argument
and rejections. Finally, we intend to create a mechanism to
indicating whether an agent must update his beliefs with
map the dialogue tables to an argumentation framework and
that information (acceptance and rejection relations).
apply some semantics to investigate the results, comparing to
The dialogue process is presented in Figure 1: the existing approaches.
References
[Amgoud and Prade, 2009] Leila Amgoud and Henri Prade.
Using arguments for making and explaining decisions. Ar-
tificial Intelligence, 173(3):413–436, 2009.
[Amgoud et al., 2002] Leila Amgoud, Nicolas Maudet, and
Simon Parsons. An argumentation-based semantics for
agent communication languages. In ECAI, volume 2,
pages 38–42, 2002.
[Carstens et al., 2015] Lucas Carstens, Xiuyi Fan, Yang
Gao, and Francesca Toni. An overview of argumentation
frameworks for decision support. In Graph Structures for
Knowledge Representation and Reasoning, pages 32–49.
Springer, 2015.
[Eğilmez et al., 2013] Sinan Eğilmez, Joao Martins, and
Joao Leite. Extending social abstract argumentation with
votes on attacks. In Theory and Applications of Formal
Figure 1: Sequence of steps in a dialogue
Argumentation, pages 16–31. Springer, 2013.
[García and Simari, 2004] Alejandro J García and
3 Argumentation Stage Guillermo R Simari. Defeasible logic programming:
To reach consensus, the agents must have compatible knowl- An argumentative approach. Theory and practice of logic
edge with each other. In the dialogue protocol, whenever an programming, 4(1+ 2):95–138, 2004.
agent sends an argument, the other agents communicate their [Muller and Hunter, 2012] Johannes Muller and Andrew
acceptance or refusal on each formula. Thus, agents having Hunter. An argumentation-based approach for decision
some formulas in ⌃ contradicting or endorsing formulas of making. In Tools with Artificial Intelligence (ICTAI), 2012
the presented argument communicate such refusal/acceptance IEEE 24th International Conference on, volume 1, pages
to the others using the speech acts Accept or Refuse. 564–571. IEEE, 2012.
Assuming that an argument <{a, a!b}, b> is proposed by [Parsons and McBurney, 2003] Simon Parsons and Peter
agent ag1 , we represent the formulas accepted (or rejected) in McBurney. Argumentation-based dialogues for agent co-
this argument as a[ag2 ,ag3 ] or a!b[ag2 ]. According to the ordination. Group Decision and Negotiation, 12(5):415–
trust level of agent ag1 and what extent the formulas accepted 439, 2003.
(or rejected) by the group, the strength of the argument can be [Ricci et al., 2009] Alessandro Ricci, Michele Piunti, Mirko
calculated with a weighted sum. Viroli, and Andrea Omicini. Environment programming in
This strength will be used to create a graph based on an cartago. In Multi-Agent Programming: Languages, Tools
abstract argumentation with weighted arguments. After map- and Applications, pages 259–288. Springer, 2009.
ping each dialogue table to a graph, we may use some se-
mantic to obtain the preferred arguments for every decision [Toni, 2014] Francesca Toni. A tutorial on assumption-based
alternative, and create a mechanism to compare all the results argumentation. Argument & Computation, 5(1):89–117,
to compute the preferred alternative among all agents. 2014.
15