Adaptive computation offloading for latency-sensitive tasks in heterogeneous edge-cloud-enabled smart warehouses using Gau-Angle FIS and AGE-MOEA-II

Industry 4.0 has introduced new development opportunities and ideological guidance to the structural reform of the supply side, as well as new requirements and challenges for the warehousing industry. Building a smart and effective warehouse system has become a crucial point for enterprises to gain a competitive advantage in market competition. As a conduit and link connecting all aspects of logistics, smart warehouses are confronted with increasing digital information and higher complexity and variability than traditional paradigms. In light of this, we establish an edge-cloud-enabled smart warehouse system and model the large-scale time-sensitive tasks based on directed acyclic graphs (DAGs). Then, by analyzing the service procedure and the cost of the devised system, an adaptive computation offloading method is proposed, which employs the Gau-Angle fuzzy inference system (FIS) for adaptive updating operators and achieves precise population screening based on the improved Pareto front modeling algorithm for large-scale many-objective optimization (AGE-MOEA-II), named ACOL-ECW. Finally, the effectiveness and superiority of ACOL-ECW are proved by comparative experiments conducted at various data scales and service scenarios.