To address the challenges posed by exponential data growth, mobile edge computing (MEC) has emerged as a key solution by decentralizing server resources from the cloud to the network edge near mobile users, thereby facilitating high-quality and low-latency service delivery. However, failures often occur in real-world mobile networks, making it crucial to consider network robustness in addition to user coverage when deploying edge servers in an MEC network. In this paper, the Geographic space-oriented Search algorithm for Edge Server Placement (ESP-GS) is proposed to optimize both of these objectives. The core idea behind ESP-GS is to leverage a technique known as “Constraint Relaxation” to downscale the edge server placement from an m-dimensional combinatorial problem to a k-dimensional continuous problem in the geographic space. This approach significantly enhances scalability, making it well-suited for large-scale network deployments. Furthermore, since the robustness of mobile edge networks lacks a widely accepted metric, a Random Failure Evaluation (RFEval) method and two corresponding metrics are designed to assess failure tolerance in practical scenarios. Extensive comparative experiments have been conducted on real-world and publicly available datasets. The results show that the ESP-GS algorithm exhibits excellent performance in both user coverage and network robustness, improving the overall network performance by 8% to 27% compared to other benchmark algorithms.

Software developers can only obtain a very small amount of information from the individual failure-causing inputs, which makes debugging difficult. Therefore, it is necessary to explore additional failure-causing inputs (failure regions) using the known failure-causing inputs. In order to accurately and efficiently identify the failure region, we propose a novel two-stage search algorithm, TS-FRI. In the initial exploration stage, a round-robin search identifies several boundary failure-causing points, and the failure region's centroid is estimated. During the main search stage, the boundary failure-causing points are identified through iterative division of the input domain with an equally sized partitioning strategy. This results in the boundary points being as dispersed as possible around the failure-region boundary, with the polytope formed by the points approximating the failure region (e.g., a polygon in two dimensions). The proposed algorithm is validated through simulation and empirical analysis: The experimental results show that the TS-FRI accuracy is at least comparable to the best accuracy of the compared three algorithms, and can be ten times better. In addition, TS-FRI only takes a quarter of the computation time and half the failure-validation cost of the other algorithms.

Summary When selecting services from a cloud center to build applications, the quality of service (QoS) is an important nonfunctional attribute to be considered. However, in actual application scenarios, the QoS details for many services may not be available. This has led to a situation where prediction of the missing QoS records for services has become a key problem for service selection. This article presents a selective ensemble learning (SEL) framework for prefilling‐based matrix factorization (PFMF) predictors. In each PFMF predictor, the improved collaborative filtering is defined by examining the stability of the QoS records when measuring the similarity of users (or services), and then used to prefill empty records in the initial QoS matrix. To ensure the diversity of the basic PFMF predictors, various prefilled QoS matrices are constructed for the matrix factorization. In this process, different reference weights are assigned to the original and the prefilled QoS records. Finally, particle swarm optimization is used to set the ensemble weights for the basic PFMF predictors. The proposed SEL on PFMF (SEL‐PFMF) algorithm is validated on a public dataset, where its prediction performance outperforms the state‐of‐the‐art algorithms, and also shows good stability.

Abstract In recent years, deep learning (DL) systems are increasingly used in the safety‐critical fields such as autonomous driving, medical diagnosis, and financial service. Although these systems have demonstrated an outstanding performance in enhancing the accuracy of decision‐making, they pose significant challenges to the trustworthiness due to their limited interpretability and inherent uncertainty. Adaptive random testing (ART) has been proved as an effective approach for ensuring the reliability of DL systems. However, existing ART methods for DL systems incur a heavy overhead in test case selection due to the computation of distances. To address this issue, we propose a lightweight adaptive random testing (Lw‐ARTDL) method for DL systems. In our improved algorithm, we employ the K‐Means technique to divide the entire test suite into several subsets. Then, for a candidate test case, we only calculate distances between it and the test cases within the category to which it belongs. This partition strategy ensures that the selected test cases are more representative while significantly reducing the computational cost. To validate the proposed algorithm, the comparison experiments between Lw‐ARTDL and the original ARTDL algorithm are conducted on two typical DL systems. The experimental results show that Lw‐ARTDL significantly reduces the overhead of failure detection, and exhibits stronger failure detection capability compared to ARTDL in most similarity metrics.

Adaptive random testing (ART) achieves better failure‐detection effectiveness than random testing due to its even spreading of test cases. ART by random partitioning (RP‐ART) is a lightweight method, but its advantage over random testing is relatively low. Although iterative partition testing (IPT) method has good performance for detecting failures in a block pattern, it loses randomness during the test case generation. To overcome the shortcomings of the above two algorithms, a new algorithm named ART by flexible partitioning (FP‐ART) is proposed. In the FP‐ART, a set of random candidates is used to select an appropriate test case by considering their boundary distance. Accordingly, the corresponding sub‐domain is also partitioned by the new test case. Based on this kind of flexible partitioning, the randomness of test case selection can be guaranteed and the spatial distribution of test cases is even more diverse. According to the results in simulation and empirical experiments, FP‐ART demonstrates better failure‐detection effectiveness than RP‐ART and is more suitable to detect the failures in strip patterns than the IPT method. Meanwhile, its failure‐detection ability is much stronger than that of fixed‐size‐candidate‐set ART in the cases of a relatively high failure rate.