Saudi Cultural Missions Theses & Dissertations
Permanent URI for this communityhttps://drepo.sdl.edu.sa/handle/20.500.14154/10
Browse
4 results
Search Results
Item Restricted ELEMENT AND EVENT-BASED TEST SUITE REDUCTION FOR ANDROID TEST SUITES GENERATED BY REINFORCEMENT LEARNING(University of North Texas, 2024-06-12) Alenzi, Abdullah Sawdi M; Bryce, ReneeAndroid stands as one of the most popular operating systems on a global scale. Given the popularity and the tremendous use of Android apps and the necessity of developing robust and reliable apps, it is crucial to create efficient and effective testing tools while addressing real-world time and budget constraints. Recently, automated test generation with Reinforcement Learning algorithms have shown promise, but there is room for improvement as these algorithms often produce test suites with redundant coverage. Fine tuning parameters of RL algorithms may assist, however, this comes with trade-offs and requires time consuming and careful consideration of the characteristics of the application under test and its environment. Therefore, devising cost-effective tools and techniques is imperative to mitigate this redundancy. Instead of exploring parameters of RL algorithms, we looked at minimizing test suites that have already been generated based on SARSA algorithms. In this dissertation, we hypothesize that there is room for improvement by introducing novel hybrid approaches that combine SARSA-generated test suites with greedy reduction algorithms following the principle of HGS approach. In addition, we apply an empirical study on Android test suites that reveals the value of these new hybrid methods. Our novel approaches focus on post-processing test suites by applying greedy reduction algorithms. To reduce Android test suites, we utilize different coverage criteria including Event-Based Criterion (EBC), Element-Based Criterion (ELBC), and Combinatorial-Based Sequences Criteria (CBSC) that follow the principle of combinatorial testing to generate sequences of events and elements. The proposed criteria effectively decreased the test suites generated by SARSA and revealed a high performance in maintaining code coverage. These findings suggest that test suite reduction using these criteria is particularly well suited for SARSA-generated test suites of Android apps.51 0Item Restricted Sample Efficient Reinforcement Learning with Partial Dynamics Knowledge(Massachusetts Institute of Technology, 2024-05-17) Alharbi, Meshal; Roozbehani, MardavijThe problem of sample complexity of online reinforcement learning is often studied in the literature without taking into account any partial knowledge about the system dynamics that could potentially accelerate the learning process. In this thesis, we study the sample complexity of online Q-learning methods when some prior knowledge about the dynamics is available or can be learned efficiently. We focus on systems that evolve according to an additive disturbance model where the underlying dynamics are described by a deterministic function of states and actions, along with an unknown additive disturbance that is independent of states and actions. In the setting of finite Markov decision processes, we present an optimistic Q-learning algorithm that achieves Õ(√T) regret without polynomial dependency on the number of states and actions under perfect knowledge of the dynamics function. This is in contrast to the typical Õ(√SAT) regret for existing Q-learning methods. Further, if only a noisy estimate of the dynamics function is available, our method can learn an approximately optimal policy in a number of samples that is independent of the cardinalities of state and action spaces. The sub-optimality gap depends on the approximation error of the noisy estimate, as well as the Lipschitz constant of the corresponding optimal value function. Our approach does not require modeling of the transition probabilities and enjoys the same memory complexity as model-free methods.39 0Item Restricted Artificial Intelligence Applied To Cybersecurity And Health Care(NDSU, 2023-09-21) Alenezi, Rafa; Ludwig, SimoneNowadays, artificial intelligence is being considered a potential solution for various problems, including classification and regression optimization, in different fields such as science, technology, and humanities. It can also be applied in areas such as cybersecurity and healthcare. With the increasing complexity and impact of cybersecurity threats, it is essential to develop mechanisms for detecting new types of attacks. Hackers often target the Domain Name Server (DNS) component of a network architecture, which stores information about IP addresses and associated domain names, to gain access to a server or compromise network connectivity. Machine learning techniques can be used not only for cyber threat detection but also for other applications in various fields. In this dissertation, the first research investigates the use of classification models, including Random Forest classifiers, Keras Sequential algorithms, and XGBoost classification, for detecting attacks. Additionally, Tree, Deep, and Kernel of Shapley Additive Explanations (SHAP) can be used to interpret the results of these models. The second research focuses on detecting DNS attacks using appropriate classifiers to enable quick and effective responses. In the medical field, there is a growing trend of using algorithms to identify diseases, particularly in medical imaging. Deep learning models have been developed to detect pneumonia, but their accuracy is not always optimal and they require large data sets for training. Two studies were conducted to develop more accurate detection models for pneumonia in chest X-ray images. The third study developed a model based on Reinforcement Learning (RL) with Convolutional Neural Network (CNN) and showed improved accuracy values. The fourth study used Teaching Learning Based Optimization (TLBO) with Convolutional Neural Network (CNN) to improve pneumonia detection accuracy, which resulted in high-level accuracy rates. Overall, all these studies provide insights into the potential of artificial intelligence in improving disease detection and cyber treat detection.33 0Item Restricted Energy Efficient D2D Communications Underlay Future Wireless Networks(University of Exeter, 2023-05-09) Alenezi, Sami Mohammed L; Min, Geyong; Luo, ChunboFrom the first generation of mobile networks to the present, the demand for more network bandwidth and energy has grown significantly as a result of the growth in users and applications. In the future, there will be billions of heterogeneous connected devices requiring high-quality network services. The demands of these cellular users are difficult to be satisfied by the technologies currently available particularly due to the limited spectrum resources. Device to Device (D2D) communication is a potential strategy for improving device performance by allowing direct communication between user pairs that are close to each other. Reducing network latency, decreasing energy consumption, increasing throughput, and improving coverage area are potential advantages of using D2D communications. However, key problems may arise when operating D2D communications in cellular networks to directly or indirectly affect energy and spectrum efficiency, for example, the interference problems between D2D devices, the interference between D2D devices and cellular devices, device discovery problems, and mode selection problems. Although traditional techniques have been proposed to solve such problems, device position, power transmission, and channel conditions are typically dynamic, particularly in the future dynamic cellular network environment. Because traditional optimisation techniques are facing increasing difficulty in rapidly changing environments, machine learning techniques become a promising tool for effective resource allocation and interference management. From this standpoint, this thesis aims to propose methods based on machine learning in order to increase the energy efficiency of D2D-assisted cellular networks. The contributions from the Machine Learning view are that the state space, action space and reward function are defined in a distributed and centralised manner to further specify the problem and use the reinforcement learning-based method to maximise energy efficiency. To be more specific, the key contributions in this thesis are listed as follows: - Few studies have been conducted to investigate the impact of user mobility on energy and spectrum efficiency of D2D communications. The effect of user mobility on the energy efficiency of D2D communications in the high-speed scenario has not been thoroughly studied especially in the state-of-the-art research in which user speed is considered very low. Thus, more research is needed to explain how D2D performance could be improved in dynamic scenarios. This thesis investigates 1) the impact of mobility on D2D communication in order to better understand the operational efficiency of next-generation cellular network-assisted D2D technologies; 2) the potential of Machine Learning (ML) algorithms to mitigate the negative impact of unpredictable user mobility; and 3) the performance gain of the proposed methods over other ML and more traditional methods. - The thesis further studies the energy efficiency of D2D communications in cellular networks. In particular, it proposes a centralised power control algorithm based on reinforcement learning to optimise energy usage while minimising interference to cellular users in order to maintain the Quality of Service (QoS). The centralised power control algorithm is hosted at the base station. Compared to the benchmark algorithms, simulation results show that the proposed method can effectively increase system energy efficiency while maintaining cellular user QoS. - Moreover, to optimise resource allocation and improve energy efficiency, the thesis proposes a Proximal Policy Optimisation (PPO)-based joint channel selection and power allocation scheme based on the Markov Decision Process (MDP). Channel selection and power allocation are jointly considered with the aim to maximise the overall energy efficiency of the network while guaranteeing the minimum requirement of QoS. Extensive simulation experiments have been carried out to validate the effectiveness of the proposed method. In terms of energy efficiency and training time, the results show that the proposed method outperforms other existing algorithms.21 0