Joint resource allocation and privacy protection for MEC task offloading in industrial Internet

The rapid development of the industrial Internet has led to significant advancements in manufacturing information. To address the limitations of traditional cloud computing architectures, such as high latency, inefficiency, and reduced flexibility, multi-access edge computing (MEC) has emerged as a promising paradigm. By enabling task offloading to edge or cloud servers, MEC can significantly reduce the task response latency and mitigate data privacy risk. However, despite the potential of MEC, many existing solutions still face challenges in balancing latency, resource allocation, and privacy protection in real-world scenarios. To address these challenges, this paper establishes a novel joint resource allocation and privacy protection offloading optimization model for multi-network transmission tasks. The model integrates offloading ratio, resource allocation, and network selection into a unified framework to minimize latency while addressing critical factors such as terminal security, latency sensitivity, and resource constraints. We formulate this problem and prove its NP-hardness before decomposing it into two subproblems. The upper layer solves the optimal offloading strategy, while the lower layer determines a reasonable resource allocation strategy based on the optimal upper-layer solution. In response, we propose a dual-layer offloading algorithm based on an adaptive parameter adjustment of the dung beetle optimizer with Bernoulli mapping. Experimental results demonstrate that our offloading model effectively enhances privacy protection, reduces latency, and optimizes overall costs. Compared with existing offloading algorithms, our algorithm shows superior convergence and offloading performance, reducing weighted costs by at least 16.91%.