A5080164144- Advanced MIMO Systems Optimization
- Age of Information Optimization
- Wireless Networks and Protocols
- IoT and Edge/Fog Computing
- Context-Aware Activity Recognition Systems
- Online Learning and Analytics
- Energy Harvesting in Wireless Networks
Hanyang University
2021-2024
We consider a multi-user multi-server mobile edge computing (MEC) system, in which users arrive on network randomly over time and generate computation tasks, will be computed either locally their own devices or offloaded to one of the MEC servers. Under such dynamic environment, we propose novel task offloading policy based hybrid online-offline learning, can efficiently reduce overall delay energy consumption only with information available at nearest servers from each user. provide...
We consider a multichannel random access system in which each user accesses single channel at time slot to communicate with an point (AP). Users arrive the and be activated for certain period of slots then disappear from system. Under such dynamic network environment, we propose distributed protocol based on multi-agent reinforcement learning (RL) improve both throughput fairness between active users. Unlike previous approaches adjusting probabilities slot, proposed RL algorithm...
A multichannel random access system is considered in which each user accesses a single channel among multiple orthogonal channels to communicate with an point (AP). Users arrive the at and be activated for certain period of time slots then disappear from system. Under such dynamic network environment, we propose distributed protocol based on multi-agent reinforcement learning (RL) improve both throughput fairness between users. Unlike previous approaches adjusting probabilities slot,...
We consider a multichannel random access system in which each user accesses single channel at time slot to communicate with an point (AP). Users arrive the and be activated for certain period of slots then disappear from system. Under such dynamic network environment, we propose distributed protocol based on multi-agent reinforcement learning (RL) improve both throughput fairness between active users. Unlike previous approaches adjusting probabilities slot, proposed RL algorithm...