COME-UP: Computation Offloading in Mobile Edge Computing with LSTM Based User Direction Prediction

• DOI: 10.3390/app12073312
• Publication Date: 2022-03-25T04:05:18Z
• Authors: Sardar Khaliq uz ...; Ali Imran Jehangiri; Tahir Maqsood; Arif Iqbal Umar; Muhammad Amir Khan; Noor Zaman Jhanjhi; Mohammad Shorfuzz...; Mehedi Masud
• Keywords: Mobile Edge Computing; Computation offloading

In mobile edge computing (MEC), devices limited to computation and memory resources offload compute-intensive tasks nearby servers. User movement causes frequent handovers in 5G urban networks. The resultant delays task execution due unknown user position base station lead increased energy consumption resource wastage. current MEC offloading solutions separate from mobility. For offloading, techniques that predict the user’s future location do not consider direction. We propose a framework termed COME-UP Computation Offloading with Long-short term (LSTM) based direction prediction. nature of mobility data is nonlinear leads time series prediction problem. LSTM considers previous features, such as location, velocity, direction, input feed-forward mechanism train learning model next location. proposed architecture also uses fitness function calculate priority weights for selecting an optimum server on latency, energy, load. simulation results show latency are lower than baseline techniques, while utilization enhanced.