The topic of this project arises naturally both as a way to meet the strategic needs of the region and to take advantage
from the opportunity posed by China as it becomes the world?s largest tourist emitter today. China is a country that
maintains good relations with Portugal and is developing a middle class with the type of tourist that fits the profile
of the target tourist of the northern region of Portugal. In addition, Portugal has to offer many of those that are the
interests and items demanded by the Chinese tourist (language, perceptions of safety in travel, cost or Government
control (e.g. obtaining a visa), etc.) [1, 2].
GrouPlanner will cover the areas of Business Tourism, City and Short Breaks, Gastronomy, Nature Tourism, Religion
Tourism, Heritage, Culture, Health and Well-being, etc. Turism of Porto and Northern Portugal, E.R. has identified
many POIs and routes. North region of Portugal has 4 world heritage sites recognized by UNESCO: Historic Centers
of Porto and Guimarães cities; Alto Douro Wine Region and Prehistoric Rock Art Sites in the Côa Valley.
Several works have been published referring research on intelligent tourism recommender systems [3], with some
emphasis on Web-based recommenders [4, 5], mobile devices [6, 7], and the use of context awareness [8]. According
to [9], travel recommendation systems aim to match the characteristics of tourism and leisure resources and attractions
with the user needs. Systems like TOURSPLAN, developed by members of the research team involved in GrouPlanner
proposal [10-12], uses optimization techniques aimed to define and adapt to the visitant?s profile a visit plan combining
the most adequate tourism products and points of interest. Some artificial intelligence techniques have been applied to
intelligent tourism recommender systems, such as: intelligent agents [13], automated scheduling and route generation
[14], heuristics [15], clustering [16], fuzzy logic and Bayesian networks [17], rule-based systems [18], ontologies [19],
etc.
The vast majority of existing intelligent tourism recommender systems are targeted at individuals, however tourism or
leisure activities are mostly performed by groups [20]. Thus, about a decade ago interest began shifting to investigate
recommendation-oriented mechanisms for groups [21-23].
However, the amount of work done on intelligent tourism recommender systems (or Group Recommender Systems
- GRS) is quite limited [24]. [25] presented a paper proposing a methodology that seeks to make recommendations
using a consensus mechanism. In the accomplishment of this work they used techniques already widely discussed in
the literature of Group Decision Making, namely in the development of Group Decision Support Systems. According
to [26], GRS are one of the most challenging topics today in the field of recommender systems. GRS draw heavily
from existing knowledge in the literature of intelligent tourism recommender systems and Group Decision Support
Systems, two topics in which the members of the GrouPlanner project are experts.
In order to support groups in decision making many approaches have been discussed in the literature [27-30]. There
is an exponential advance and an increasingly high level of complexity [31-33]. Multi-agent systems have been used
to better represent the participants [34, 35]. The use of intelligent agents to represent participants allows participants
to be represented not only in relation to their preferences but also in accordance with their intentions [36], personality
[37], interests and goals/objectives [38]. In addition, the autonomy inherent to intelligent agents also allows them to
be able to perform more complex tasks, such as obtaining information from the environment, being proactive, sharing
information and resources, cooperating and coordinating their activities, etc. [39]. Also under the topic of intelligent
tourism recommender systems we can find approaches that include multi-agent systems [40-43]. However, several
approaches, such as [13, 40, 44] make use of agents to represent: POIs, system components and experts in certain
topics (flights, hotels or attractions) using fairly simple types of communications. There are no approaches concerning
multi-agent systems capable of representing each of the elements of the group, dealing with their preferences,
intentions, relationships with other elements of the group, etc.
As mentioned there are several mechanisms under the topic of intelligent tourism recommender systems with goal
of proposing solutions to tourists. However, these mechanisms do not allow to justify options during the "negotiation"
process. The use of argumentation-based negotiation models permits overcoming this problem. The main idea of
argumentation-based negotiation is the ability to support offers with justifications and explanations, which play a key
role in negotiation settings [45]. So, it allows the participants in the negotiation not only to exchange offers, but also
reasons and justifications that support these offers in order to mutually influence their preference relations on the
set of offers, and consequently the outcome of the dialogue [45]. Clearly, planning a trip, deciding points to visit, or
simply choosing a restaurant to go to dinner in group involves various types of dialogue (in a physical/real context).
According to the types of dialogues identified by [46] we can easily perceive that in this kind of decision can be put
into practice types such as: persuasion, negotiation, inquiry, information seeking and deliberation. Dialogue systems
have also been studied by computer science scientists [47-49]. Modeling dialogues using argumentation is a complex
task and the approaches in the literature are still very much oriented towards one-to-one dialogue [50-53]. Even more
unprecedent is a model that allows real agents and participants to be included in the same process.