Planejamento de rotas veiculares e otimização de mobilidade urbana utilizando algoritmo bioinspirado e paralelo

• Ano de defesa: 2022
• Autor(a) principal: Paz, Fillipe Almeida
• Orientador(a): Matos Júnior, Rubens de Souza
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Programa de Pós-Graduação: Pós-Graduação em Ciência da Computação
• Palavras-chave em Português: Algoritmos paralelos; Sistemas inteligentes de veículos rodoviários; Telecomunicações; Inteligência artificial; Otimização; Computação paralela; Intelligent transport systems (ITS); Mobilidade urbana; 5G
• Palavras-chave em Inglês: Optimization; Parallel computing; Urban mobility
• Área do conhecimento CNPq: CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
• Link de acesso: https://ri.ufs.br/jspui/handle/riufs/18320

The concentration of the world’s population in urban areas has posed challenges to the quality of life of citizens. Traffic jams, increased air pollution, waste of natural resources and reduced productivity of citizens are some examples of damage resulting from poor mobility in cities. Simultaneously, technologies applied to cities can help mitigate these effects, improving society’s quality of life and producing a more environmentally and economically sustainable environment. Processes and technologies for data collection, analysis and transmission make up the ecosystem of smart cities and can help in decision-making in order to optimize resources. However, although the amount of data regarding the situation of urban traffic is considerable, the dynamic planning of vehicular routes that takes into account a large amount of roads, vehicles and calculations on real-time data is limited due to the computing time involved. In this context, this work aims to experimentally evaluate the impact of dynamic vehicular routing on metrics related to urban mobility and computing time. Heuristic-based, classical and bio-inspired algorithms were evaluated under different road flow conditions, with emphasis on PPUMO (Parallel and Pheromone-based Urban Mobility Optmization), an algorithm for vehicular routing proposed in this work. It was evidenced in the Results of this work that PPUMO was able to produce the best results among the treatments when analyzed in relation to the metrics Timeloss (reduction of up to 27.8%, on average); Trip Duration (up to 28.4% reduction on average); Jams (up to 73.9% reduction on average) for high load conditions on the traffic system. It was also possible to verify that the PPUMO presented a route replanning time 10000 times smaller compared to the serial algorithms. Especially with regard to vehicle travel time, as it was the best choice in 80% of the evaluated scenarios. Finally, with regard to route length, PPUMO achieved lower or close to lower results compared to the base case and distanced itself from the approaches that performed route replanning by up to 32.4%, on average. This work contributes to the identification of strengths and weaknesses associated with the use of dynamic routing based on deterministic or probabilistic, classical or bioinspired algorithms. Here is also proposed a dynamic vehicular routing architecture that foresees the use of 5G V2X networks for data transmission in an Intelligent Transport System. Finally, it is intended that the implementations of this work serve as a software artifact for the development of solutions for the maximization of urban mobility.