http://repositorio.ufc.br/handle/riufc/24011 Sat, 12 Sep 2026 17:59:22 GMT 2026-09-12T17:59:22Z http://repositorio.ufc.br/handle/riufc/87650 Título: Problema de formação de múltiplos times com múltiplas habilidades: uma avaliação experimental da heurística baseada em algoritmos genéticos Autor(es): Santos, Pedro Henrique Maia dos Abstract: The Multiple Team Formation with Multiple Skills Problem (MMTFP) consists of distributing a set of individuals, each possessing different skills and time allocation capacities, among multiple teams, so as to meet predefined skill requirements and maximize the harmony among team members. This work aims to evaluate the performance of the Genetic Algorithm-based heuristic proposed by Viana (2018) for the MMTFP, comparing its solutions with the optimal values obtained by the Integer Linear Programming formulation of Figueiredo (2021). Since optimal values were not available at the time of its publication, Viana (2018) could not calculate the optimality gap of her solutions for the MMTFP instances. This work fills that gap by obtaining the optimal values using the CPLEX 12.8 solver and calculating, for the first time, the relative optimality gap for the 135 synthetic instances generated by Viana (2018), organized into three categories according to the percentage of individuals with multiple skills, three classes according to the granularity of time dedication fractions, and three groups according to the density of positive relations in the sociometric matrix. The results indicate that the factor most influencing the heuristic performance is the density of positive relations in the sociometric matrix, with an average gap of 10.25% for the group with the lowest density and 2.87% for the group with the highest density. The variation in the percentage of individuals with multiple skills produced no clear trend on solution quality, indicating robustness of the heuristic to this problem dimension. The overall average gap was 6.46%, with a minimum of 0.50% and a maximum of 17.91%. Tipo: TCC Thu, 01 Jan 2026 00:00:00 GMT http://repositorio.ufc.br/handle/riufc/87650 2026-01-01T00:00:00Z http://repositorio.ufc.br/handle/riufc/87634 Título: Aplicação de Programação por Conjunto de Respostas e Metaheurísticas para o Problema de Alocação de Voluntários com Preferências Autor(es): Dutra, Francisco Rian de Oliveira Abstract: The Volunteer Assignment Problem, a variant of the Generalized Assignment Problem classified as NP-hard, poses significant computational challenges, making exhaustive search computationally prohibitive. This work evaluates three solution strategies: exact search using Answer Set Programming (ASP) and two hybrid approaches that use, respectively, a greedy heuristic and ASP to generate initial solutions, which are subsequently refined by the metaheuristics Simulated Annealing (SA) and Genetic Algorithm (GA). The objective is to maximize a function based on volunteer preferences. The results show that the exact method, when used in isolation, reaches the time limit without proving optimality, whereas the hybrid strategies produce high-quality solutions within practical computational times. Notably, the combination of an initial solution generated by the exact solver (ASP) with refinement by the Genetic Algorithm achieved the best overall performance, effectively addressing the complexity of the problem and outperforming the constructive heuristic in maximizing allocation quality. Tipo: TCC Thu, 01 Jan 2026 00:00:00 GMT http://repositorio.ufc.br/handle/riufc/87634 2026-01-01T00:00:00Z http://repositorio.ufc.br/handle/riufc/87632 Título: Aplicação de programação por conjunto de respostas e metaheurísticas para o problema de alocação de voluntários com preferências Autor(es): Dutra, Francisco Rian de Oliveira Abstract: The Volunteer Assignment Problem, a variant of the Generalized Assignment Problem classified as NP-hard, poses significant computational challenges in which exhaustive search becomes computationally prohibitive. This work evaluates three solution strategies: exact search via Answer Set Programming (ASP) and two hybrid approaches that use, respectively, a greedy heuristic and ASP itself to generate initial solutions, subsequently refined by the metaheuristics simulated annealing (SA) and genetic algorithm (GA). The objective is to maximize a function based on volunteer preferences. Results show that the exact method alone hits the time limit without proving optimality, while hybrid strategies deliver high-quality solutions within practical time constraints. Notably, the initialization generated by the exact solver (ASP) combined with refinement by the Genetic Algorithm achieved the best overall performance, overcoming problem complexity and outperforming the constructive heuristic in maximizing allocation quality. Tipo: TCC Thu, 01 Jan 2026 00:00:00 GMT http://repositorio.ufc.br/handle/riufc/87632 2026-01-01T00:00:00Z http://repositorio.ufc.br/handle/riufc/87630 Título: Uso de inteligência artificial em tecnologias assistivas para estudantes com TEA: personalização de ensino e monitoramento de progresso Autor(es): Sousa, Lara Amanny Ramos de Abstract: This study investigated the role of Artificial Intelligence (AI) applied to assistive technologies for students with Autism Spectrum Disorder (ASD), with an emphasis on personalized teaching and monitoring educational progress. This is an integrative review with a qualitative approach that analyzed 11 articles published between 2018 and 2025, selected for their direct relevance to the application of AI in inclusive contexts. The results show that AI contributes to adapting content to individual needs, offering real-time feedback, and generating performance indicators, favoring greater engagement, autonomy, and participation of students. The study stands out for proposing practical actions that are feasible in public schools, such as the use of free Augmentative and Alternative Communication (AAC) applications, the definition of gradual communication goals and performance frameworks, as well as suggesting simple metrics for monitoring (response time, number of interactions, and engagement levels). It concludes that the incorporation of AI into assistive technologies represents a significant advance for the consolidation of inclusive education, provided it is accompanied by scientific rigor, teacher mediation, and ethical data governance, in accordance with the General Data Protection Law (LGPD). Tipo: TCC Thu, 01 Jan 2026 00:00:00 GMT http://repositorio.ufc.br/handle/riufc/87630 2026-01-01T00:00:00Z