A certificação CERTICS foi desenvolvida para ser um instrumento de política pública que busca contribuir para o desenvolvimento nacional sustentável e pode apoiar as empresas nacionais de software na evolução necessária para se tornarem mais competitivas frente aos softwares estrangeiros. No entanto, esta certificação, assim como outras, requer investimento de profissionais e recursos financeiros, o que é um problema notadamente nas pequenas empresas de software. Este trabalho tem o objetivo de apresentar o GARREC, Guia para Atendimento dos Requisitos dos Resultados Esperados da CERTICS, que é uma ferramenta desenvolvida para apoiar no processo da certificação CERTICS, atuando em complemento à documentação existente. O GARREC foi construído visando facilitar o entendimento dos conceitos da CERTICS e no atendimento dos resultados esperados por meio de proposição de evidências, considerando cenários de pequenas empresas. Assim, o GARREC contribuirá para reduzir o investimento necessário para a certificação. O método de pesquisa adotado envolveu a análise do Modelo de Referência para Avaliação da CERTICS e o detalhamento dos Requisitos Específicos dos seus Resultados Esperados e, para estes foram propostas evidências para atendimento classificadas por relevância. Desta forma, todos os aspectos avaliados são considerados, garantindo qualidade de cobertura do atendimento aos requisitos da certificação. Para a avaliação do GARREC foi realizado um experimento no qual os participantes o utilizaram para atender a resultados esperados predeterminados e responderam a uma pesquisa. Participaram do experimento três empresas com diferentes níveis de conhecimento da CERTICS, uma empresa certificada, uma em processo de certificação e uma sem conhecimento anterior. A partir dos resultados coletados da pesquisa de avaliação, o GARREC atinge os seus objetivos de auxiliar no entendimento e no atendimento dos requisitos da certificação CERTICS, com 91,3% de aceitação aos itens de efetividade e 97,5% referente aos itens de aplicabilidade. Uma validação mais ampla em campo ainda se faz necessária para uma avaliação mais consistente da ferramenta. / The CERTICS certification was developed to be a public policy tool that seeks to contribute to sustainable national development and it can support national software companies in the evolution required to become more competitive compared to the foreign software. However, this certification, as well as others, requires professional investment and financial resources, which is usually a problem for small software companies. This work aims to present GARREC, Guide for Meeting the Requirements of Results Expected from CERTICS, which is a tool developed to support the understanding and obtaining of the CERTICS certification, working in addition to the existing documentation. GARREC was built to facilitate the understanding of the CERTICS’ concepts and in meeting the expected results through evidence proposition considering small business scenarios.Therefore, GARREC will contribute to reducing the investment required for certification. The research method involved the analysis of the Reference Model for Evaluation of CERTICS and detailing of the Specific Requirements of its Expected Results, and for these, evidence was presented to meet them, classified by relevance. In this way all evaluated aspects are considered, guaranteeing quality of coverage of the attendance to the certification requirements. For the GARREC evaluation, an experiment was carried out in which the participants used it to meet predetermined expected results and answered to a survey. Three companies with different levels of knowledge of CERTICS, a certified company, one in the process of certification and one without previous knowledge participated in the experiment. Based on the results of the evaluation survey, GARREC achieves its objectives of assisting in the understanding and fulfillment of CERTICS certification requirements, with 91.3% acceptance of the items referring to Effectiveness and, 97.5% acceptance of the related items Applicability. Further validation in the field is still necessary for a more consistent evaluation of the tool.

Software prediction is a complex area as there are no accurate models to represent reliability throughout the use of software, unlike hardware reliability. In the context of the software reliability of on-board train systems, ensuring good software reliability over time is all the more critical given the current density of rail traffic and the risk of accidents resulting from a software malfunction. This thesis proposes to use soft computing methods and historical failure data to predict the software reliability of on-board train systems. For this purpose, four machine learning models (Multi-Layer Perceptron, Imperialist Competitive Algorithm Multi-Layer Perceptron, Long Short-Term Memory Network and Convolutional Neural Network) are compared to determine which has the best prediction performance. We also study the impact of having one or more features represented in the dataset used to train the models. The performance of the different models is evaluated using the Mean Absolute Error, Mean Squared Error, Root Mean Squared Error and the R Squared. The report shows that the Long Short-Term Memory Network is the best performing model on the data used for this project. It also shows that datasets with a single feature achieve better prediction. However, the small amount of data available to conduct the experiments in this project may have impacted the results obtained, which makes further investigations necessary. / Att förutsäga programvara är ett komplext område eftersom det inte finns några exakta modeller för att representera tillförlitligheten under hela programvaruanvändningen, till skillnad från hårdvarutillförlitlighet. När det gäller programvarans tillförlitlighet i fordonsbaserade tågsystem är det ännu viktigare att säkerställa en god tillförlitlighet över tiden med tanke på den nuvarande tätheten i järnvägstrafiken och risken för olyckor till följd av ett programvarufel. I den här avhandlingen föreslås att man använder mjuka beräkningsmetoder och historiska data om fel för att förutsäga programvarans tillförlitlighet i fordonsbaserade tågsystem. För detta ändamål jämförs fyra modeller för maskininlärning (Multi-Layer Perceptron, Imperialist Competitive Algorithm Mult-iLayer Perceptron, Long Short-Term Memory Network och Convolutional Neural Network) för att fastställa vilken som har den bästa förutsägelseprestandan. Vi undersöker också effekten av att ha en eller flera funktioner representerade i den datamängd som används för att träna modellerna. De olika modellernas prestanda utvärderas med hjälp av medelabsolut fel, medelkvadratfel, rotmedelkvadratfel och R-kvadrat. Rapporten visar att Long Short-Term Memory Network är den modell som ger bäst resultat på de data som använts för detta projekt. Den visar också att dataset med en enda funktion ger bättre förutsägelser. Den lilla mängd data som fanns tillgänglig för att genomföra experimenten i detta projekt kan dock ha påverkat de erhållna resultaten, vilket gör att ytterligare undersökningar är nödvändiga.