This paper investigates the stability analysis of a class of fractional-order bidirectional associative memory neural networks (FOBAMNNs) with S-type distributed delays, which incorporating both discrete and distributed delays through a Lebesgue-Stieltjes integral. Leveraging the properties of Riemann-Liouville fractional-order derivatives and integrals, along with the additivity of integration intervals and initial conditions, we transform fractional-order integrals of the state function involving S-type distributed delays into equivalent integrals without such delays. Utilizing the theory of contractive mapping and the Bellman-Gronwall inequality, the finite-time stability (FTS) and global Mittag-Leffler stability (GLMS) criteria are established when certain conditions were satisfied. Moreover, the two distinct types of stability, FTS and GMLS, in FOBAMNNs are governed by appropriate parameters, complementing each other and laying a new foundation for advancing research and practical applications in neural networks. Furthermore, the validity and practical applicability of the theoretical results are demonstrated through two illustrative numerical simulation examples.

Crowdsensing Data Trading (CDT) has emerged as a novel data trading paradigm, where market stability is crucial during the transaction matching process. However, most existing CDT systems usually assume that the preferences of both parties are known and the third-party trading platform is trustworthy, which is impractical in real-world scenarios and leads to significant challenges in reliability and privacy preservation. To address these challenges, we propose a Privacy-Preserving and Stable Data Trading for Unknown Market based on Blockchain and Bilateral Reputation (PPSDT-UMBBR) scheme in the decentralized CDT system. First, a privacy-preserving bilateral preference initialization method is designed to achieve the initial matching of buyers and sellers without exposing their location and attribute privacy. Then, a stable matching method based on dynamic bilateral preference updating is proposed, integrating Differential Privacy, Stable matching theory, and a strategy based on Asymmetric Bilateral Preferences with Multi-Armed Bandits (DPS-ABPMAB). Finally, we theoretically analyze the security and prove that the market outcome is$\delta$-stable. Furthermore, compared to other benchmark methods based on real datasets, our proposed DPS-ABPMAB algorithm improves the average accumulative reward by at least 4.22%, and reduces the average accumulative regret and the mean evaluation error rate by at least 66.86% and 7.35%, respectively.

The time-varying characteristics of user mobility, node and services connections, and edge resources in Mobile Edge Computing (MEC) scenarios pose significant challenges for designing efficient service migration strategies to enhance system performance. This paper introduces and quantifies affinity and anti-affinity metrics to evaluate the discrepancies between the resource requirements of containers and the available resources of nodes, as well as the competition level of computing resources during the migration of containers among different nodes within the Kubernetes cluster. To adapt to the complexity of the edge environment, such metrics are integrated into the reward update mechanism of the Deep Deterministic Policy Gradient (DDPG) reinforcement algorithm. Besides, the Multi-Objective Evolutionary Algorithm (MOEA) is employed to dynamically adjust the weights of various reward objectives, forming a self-adaptive online container migration strategy named MO-DDPG. Finally, we construct a real-world heterogeneous Kubernetes edge node cluster in experiments and use a public dataset to simulate multi-modal connections between mobile user trajectories and service demands. Compared to the greedy and heuristic strategies that consider only single metrics, our experiment results show that the proposed strategy improves energy efficiency and reduces latency by 21.30% and 29.49%, respectively. Moreover, the MO-DDPG improves resource utilization of the node cluster compared to the default Kubernetes scheduler.

Existing methods for decomposing monolithic applications into microservices in cloud environments primarily rely on the call relationships within itself. However, these methods are difficult to apply directly in resource-constrained and distributed edge network scenarios without considering the heterogeneity of the device. Therefore, this paper proposes a clustering method that ensembles graph structures and device features based on attention mechanism, which utilizes attention encoders to learn node embeddings and employs a spectral clustering algorithm to obtain decomposition results, optimizing the affinity and matching degree between microservices and devices. Experimental results demonstrate that the proposed method exhibits excellent performance in terms of functional independence, modularity, and adaptability of microservices.

Mobile Crowdsourcing (MCS) has become a novel paradigm for enabling data collection by worker recruitment, and the reputation plays a crucial role in achieving high-quality data. Although identity, data, and bid privacy preserving have been thoroughly investigated with the advance of blockchain technology, existing literature barely focuses on reputation privacy, which prevents malicious workers from submitting false data that could affect truth discovery for data requester. Therefore, we propose a Blockchain-Based Reputation Privacy Preserving for Quality-Aware Worker Recruitment Scheme (BRPP-QWR). First, we design a lightweight privacy preserving scheme for the whole life cycle of the worker’s reputation, which adopts sub-address retrieval technique combined with Pedersen Commitment and Compact Linkable Spontaneous Anonymous Group (CLSAG) signature to enable fast and anonymous verification of the reputation update process. Subsequently, to tackle the unknown worker recruitment problem, we propose a Reputation, Selfishness, and Quality-based Multi-Armed Bandit (RSQ-MAB) learning algorithm to select reliable and high-quality workers. Lastly, we implement a prototype system on Hyperledger Fabric to evaluate the performance of the reputation management scheme. The results indicate that the execution latency for the reputation score verification and retrieval latency can be reduced by an average of 6.30%–56.90% compared with ARMS-MCS. In addition, experimental results on both real and synthetic datasets show that the proposed RSQ-MAB algorithm achieves an increase of at least 20.05% in regard to the data requester’s total revenue and a decrease of at least 48.55% and 3.18% in regret and Multi-round Average Error (MAE), respectively, compared with other benchmark methods.