Argumentation framework. In order for the agents to consider potential mappings and the reasons for and against accepting them, we use an argumentation framework. Our framework is based on Value- based Argument Frameworks (V AFs) [3].This work is an experimental research and 4 Although the agents’ ontologies may differ, we assume that ontologies are encoded in the same language, the standard OWL (xxxx://xxx.x0.xxx/OWL/), thus eliminating the problem of integrating different ontology languages. a prototype of the framework is under development. We start with the presentation of Xxxx work [7], upon which the V AFs rely. ⟨ ⟩ ⊂ × ⟨ ⟩
Argumentation framework. Our real-time argumentation framework defines three types of individual knowledge resources that the agents can use to manage arguments: A case-base with domain-cases, that represent pre- vious problems and their solutions. Agents can use this knowledge resource to generate their po- sitions and arguments. The position of an agent represents the solution that this agent proposes, by reusing the solution applied to solve a simi- lar problem in the past. Also, agents increase their domain knowledge at the end of each real-time argumentation dialogue by adding new cases to their domain-cases case-base. A case-base with argument-cases, that store previ- ous argumentation experiences and their final out- come. Agents use this resource to select the best position and argument to put forward in a specific situation in view of how suitable a similar posi- tion or argument was in a similar real-time argu- mentation dialogue. Also, agents store the new ar- gumentation knowledge gained in each real-time agreement process, improving the agents’ argu- mentation skills. A set of argumentation-schemes, that represents stereotyped patterns of reasoning [46]. Argumentation- schemes consists of a set of premises from which agents can draw specific conclusions. In this sense, argumentation-schemes represent general rules that hold in the domain under discussion (e.g. regarding exceptional situations that force agents to select a specific type of solution). In addition, argumentation-schemes include a set of critical questions that represent possible ways of attacking the conclusion drawn from the scheme (e.g. exceptions to the rule, other sources of infor- mation that invalidate the rule, etc.). In our proposal, arguments that agents exchange are tuples of the form: ⟨ { }⟩
Argumentation framework. An argumentation framework (AF) is a pair AF = (A, R) where • A is a non-empty set of arguments, and • R ⊆ A × A is an attack relation. If (x, y) ∈ E we say that the argument x attacks argument y. An argumentation framework (AF) essentially describes an argumentation setting as a directed graph. A crucial question when faced with such an argumentation setting is what is the sensible position to take with respect to the arguments which have been “presented”, or are otherwise being assessed. There may be several viable positions to take, or occasionally no sensible position. What is sensible is not a straightforward conclusion to reach either. The/a sensible position, either way, will be some set of arguments among the arguments in question. Such a subset of the arguments is called an extension. Several semantics have been proposed. A semantic always have the same signature, as defined in the following definition. ( )∈ ( )⊆ → Definition 2.20 (Extension-based Semantic). Let A be some collection of argumenta- tion frameworks over arguments X. Xx (extension-based) semantic is a map ε : A 22A. That is, for an argumentation framework A, R A, ε A, R 2A yields a set of extensions. There are several semantics for argumentation theory. Each with properties which speak for and against it, depending on how we perceive argumentation to function, or rather how we think argumentation ought to function. Some common, and simple, candidates are listed in the following definition.
