Tracking the process of remote task execution is critical to timely process analysis by collecting the evidence of correct execution or failure, which generates a process digital twin (DT) for remote supervision. Generally, it will encounter the challenge of constrained communication, high overhead, and high traceability demand, leading to the efficient remote process tracking issue. Existing approaches can address the issue by monitoring or simulating remote task execution. Nevertheless, they do not provide a cost-effective solution, especially when unexpected situation occurs. Thus, we proposed a new cloud-edge collaboration framework for process DT generation. It addresses the efficient remote process tracking issue with a real-virtual collaborative process tracking (RVCPT) approach. The approach contains three patterns of real-virtual collaboration for tracking the entire process of task execution with a coevolution pattern, identifying unexpected situations with a discrimination pattern, and generating a process DT with a real-virtual fusion pattern. This approach can minimize tracking overhead, and meanwhile maintains high traceability, which maximizes the overall cost-effectiveness. With prototype development, case study and experimental evaluation show the applicability and performance advantage of the new cloud-edge collaboration framework in remote supervision.

As a privacy-focused distributed machine learning, federated learning can not only train models effectively but also prevent data sets from being leaked easily. However, like crowdsensing perception and other technologies, participants often lack the motivation to learn and the quality of participation is not high. Therefore, this paper mainly designs a two-stage federal learning incentive mechanism based on the Stackelberg game model under a specific model accuracy index. Firstly, we combine data quality and data quantity to construct a federal learning incentive mechanism model under specific indicators. Then, we conduct a two-stage Stackelberg game analysis on the incentive mechanism model based on the utility function construction of the server platform and data island. In the first stage, the platform server is the leader and the data island is the follower. The second stage is a Nash equilibrium game between data islands. Finally, we construct the objective function of the server platform and data island, namely utility maximization, deduce the optimal equilibrium solution of the two-stage game, and determine the optimal strategy of the platform server and data island.

DatalogMTL is an extension of Datalog with metric temporal operators that has found an increasing number of applications in recent years. Reasoning in DatalogMTL is, however, of high computational complexity, which makes reasoning in modern data-intensive applications challenging. In this paper we present a practical reasoning algorithm for the full DatalogMTL language, which we have implemented in a system called MeTeoR. Our approach effectively combines an optimised (but generally non-terminating) materialisation (a.k.a. forward chaining) procedure, which provides scalable behaviour, with an automata-based component that guarantees termination and completeness. To ensure favourable scalability of the materialisation component, we propose a novel seminaïve materialisation procedure for DatalogMTL enjoying the non-repetition property, which ensures that each rule instance will be applied at most once throughout its entire execution. Moreover, our materialisation procedure is enhanced with additional optimisations which further reduce the number of redundant computations performed during materialisation by disregarding rules as soon as it is certain that they cannot derive new facts in subsequent materialisation steps. Our extensive evaluation supports the practicality of our approach.