Web-based service providers have long been required to deliver high quality services in accordance with standards and customer requirements. Increasingly, however, providers are required to think beyond service quality and develop a deeper understanding of their customers’ Quality of Experience (QoE). Whilst models exist that assess the QoE of Web Application, significant challenges remain in defining QoE factors from a Web engineering perspective, as well as mapping between so called ‘objective’ and ‘subjective’ factors of relevance. Specifically, the following challenges are considered as general fundamental problems for assessing QoE: (1) Quantifying the relationship between QoE factors; (2) predicting QoE as well as dealing with the limited data available in relation to subjective factors; (3) optimising and controlling QoE; and (4) perceiving QoE. In response, this research presents a novel model, called QoEWA (and associated software instantiation) that integrates factors through Key Performance Indicators (KPIs) and Key Quality Indicators (KQIs). The mapping is incorporated into a correlation model that assesses QoE, in particular, that of Web Application, with a consideration of defining the factors in terms of quality requirements derived from web architecture. The data resulting from the mapping is used as input for the proposed model to develop artefacts that: quantify, predict, optimise and perceive QoE. The development of QoEWA is framed and guided by Design Science Research (DSR) approach, with the purpose of enabling providers to make more informed decisions regarding QoE and/or to optimise resources accordingly. The evaluation of the designed artefacts is based on a build-and-evaluate cycle that provides feedback and a better understanding of the utilised solutions. The key artefacts are developed and evaluated through four iterations: Iteration 1 utilises the Actual Versus-Target approach to quantify QoE, and applies statistical analysis to evaluate the outputs. Iteration 2: utilises a Machine Learning (ML) approach to predict QoE, and applies statistical tests to compare the performance of ML algorithms. Iteration 3 utilises the Multi-Objective Optimisation (MOO) approach to optimise QoE and control the balance between resources and user experience. Iteration 4 utilises the Agent-Based Modelling approach to perceive and gain insights into QoE. The design of iteration 4 is rigorously tested using verified and validated models.

Object of the study - development of an environment for training an agent controlling the actions of a robotic manipulator model. Subject of the study - application of Unity ML-Agent technology for training an agent controlling the actions of a robotic manipulator model. The purpose of the work is to develop a system for training the Unity ML-Agents agent controlling the robotic manipulator model to achieve the target object. Research methods: mathematical modeling, data analysis, experimental method. The result of the master's thesis is the successful creation of a machine learning system for controlling a robotic arm in the Unity virtual environment. / Объект исследования – разработка среды для обучения агента, управляющего действиями модели робота-манипулятора. Предмет исследования – применение технологии Unity ML-Agent для обучения агента, контролирующего действия модели робота-манипулятора. Цель работы – разработка системы для обучения агента Unity ML-Agents, управляющего моделью робота-манипулятора, достижению целевого объекта. Методы исследования: математическое моделирование, анализ данных, экспериментальный метод. Результатом магистерской работы является успешное создание системы машинного обучения для управления роботом-манипулятором в виртуальной среде Unity.