Construction of MEC Task Offloading Strategy and Risk Assessment Model Based on Multi-objective Optimization

Traditional mobile edge computing (MEC) task offloading often faces the challenge of balancing latency, energy consumption, and reliability.This study proposes a novel multi-objective optimization framework that combines a long short-term memory (LSTM) network with a Markov decision process (MDP) to achieve intelligent task offloading decisions.The LSTM component processes historical network and resource data to predict the optimal offloading strategy, while the MDP models the decision-making process under uncertainty.In addition, we develop a Bayesian network-based risk assessment model to identify potential system risks and classify them into five levels (0-IV).Experimental evaluation using three task arrival rates (0.3, 0.6, 0.9) shows that our approach achieves 6.9-56% latency reduction and a wide range of energy consumption reduction compared to four state-of-the-art algorithms (GTO, WOA, NGO, DBO), while maintaining more than 98.6% reliability.The risk assessment consistently achieves zero risk classification in all test scenarios, indicating that the system performance is robust.This study provides a comprehensive framework for MEC task offloading.Multiple objectives are optimized simultaneously while ensuring system reliability.