| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 6.28 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
O protĆ³tipo de Sistema de RecomendaĆ§Ć£o para Grupos no setor do turismo que se encontra em desenvolvimento pelo GECAD utiliza num dos seus microserviƧos um Sistema MultiAgente. No entanto, a interaĆ§Ć£o e partilha de conhecimento entre os agentes do sistema carece de melhorias que permitam a apresentaĆ§Ć£o de recomendaƧƵes cada vez mais coerentes, precisas e personalizadas. Os dados sociodemogrĆ”ficos, personalidade e preferĆŖncias turĆsticas de cada turista sĆ£o modelados num agente inteligente que o representa, com o objetivo de tornar cada agente o mais similar possĆvel ao turista que representa. Com isto, os agentes devem ter a capacidade de aprender com o conhecimento e aƧƵes dos outros agentes, assim como de partilhar dados de interaĆ§Ć£o e perfil dos turistas, de forma a melhorar e a tornar mais precisas as recomendaƧƵes enviadas pelo sistema de recomendaƧƵes do GRS, melhorando a satisfaĆ§Ć£o e experiĆŖncia dos turistas. Este documento apresenta uma anĆ”lise de valor e inclui um estudo sobre o estado da arte em tecnologia relevante e sobre conceitos e trabalhos relacionados com o projeto em desenvolvimento. Apresenta ainda uma anĆ”lise do domĆnio do problema, o desenho arquitetural e detalhes sobre a implementaĆ§Ć£o e testagem do protĆ³tipo desenvolvido. A soluĆ§Ć£o final respondeu essencialmente a todas as necessidades que se tinham proposto resolver e possibilita o crescimento do GRS sem comprometer o trabalho jĆ” efetuado. Os utilizadores passaram a ser associados a clusters com base na sua personalidade e os respetivos agentes foram melhorados para utilizar ratings e rules que lhes diziam respeito de forma a priorizar e penalizar pontos de interesse turĆsticos nas recomendaƧƵes obtidas.
The prototype of the Recommendation System for Groups (GRS) in the tourism sector that is being developed by GECAD uses a Multi-Agent System in one of its microservices. However, the interaction and knowledge sharing between the agents of the system needs improvements to allow the presentation of increasingly coherent, accurate and personalized recommendations. The socio-demographic data, personality and tourist preferences of each tourist are modeled in an intelligent agent, with the goal of making each agent as similar as possible to the tourist it represents. With this, the agents should have the ability to learn from the knowledge and actions of the other agents, as well as to share interaction data and tourist profiles, in order to improve and make more accurate the recommendations sent by the GRS recommendation system, improving tourists' satisfaction and experience. This paper presents a value analysis and includes a study of the state of the art in relevant technology and of concepts and work related to the project under development. It also presents an analysis of the problem domain, the architectural design, and details about the implementation and testing of the developed prototype. The final solution essentially met all the needs that had been proposed to solve and allows for the growth of the GRS without compromising the work already done. Users are now associated to clusters based on their personality and the respective agents were improved to use ratings and rules related to them in order to prioritize and penalize tourist points of interest in the recommendations obtained.
The prototype of the Recommendation System for Groups (GRS) in the tourism sector that is being developed by GECAD uses a Multi-Agent System in one of its microservices. However, the interaction and knowledge sharing between the agents of the system needs improvements to allow the presentation of increasingly coherent, accurate and personalized recommendations. The socio-demographic data, personality and tourist preferences of each tourist are modeled in an intelligent agent, with the goal of making each agent as similar as possible to the tourist it represents. With this, the agents should have the ability to learn from the knowledge and actions of the other agents, as well as to share interaction data and tourist profiles, in order to improve and make more accurate the recommendations sent by the GRS recommendation system, improving tourists' satisfaction and experience. This paper presents a value analysis and includes a study of the state of the art in relevant technology and of concepts and work related to the project under development. It also presents an analysis of the problem domain, the architectural design, and details about the implementation and testing of the developed prototype. The final solution essentially met all the needs that had been proposed to solve and allows for the growth of the GRS without compromising the work already done. Users are now associated to clusters based on their personality and the respective agents were improved to use ratings and rules related to them in order to prioritize and penalize tourist points of interest in the recommendations obtained.
Description
Keywords
Sistemas de RecomendaĆ§Ć£o para Grupos Turismo Pontos de Interesse Personalidade Recommendation Systems for Groups Tourism Points of Interest Personality Sistemas Multi-Agente MicroserviƧos Multi-Agente ActressMAS C# .NET Multi-Agent Systems Multi-Agent Microservices