Latency and Energy Efficient Bio-Inspired Conic Optimized and Distributed Q Learning for D2D Communication in 5G

Abstract

The next-generation communication, i.e. fifth generation (5G), will be manifesting the advertisers in near future. The Device to Device communication would be a proportion of 5G to provide communication requirements for trillions of devices connected in a significant manner to hold immense data rate. Optimization and Machine learning are said to be the most promising mechanisms for furnishing the best solutions to master the predominant aspects and explicit parameters of 5G communication. In this work, a Bio-inspired Conic Optimized and Distributed Latency Q Learning method is proposed for D2D communication in 5G with higher energy efficiency and minimum latency. Initially, a Bio-inspired Conic Particle Swarm Optimization model is to provide resource efficiency and energy efficiency by calculating the fitness function in terms of transmission power and data loss rate through updating the position and velocity to achieve optimized D2D communication. After that, the Distributed Latency Managed Q Learning model is employed for better connectivity with minimal latency which is achieved by two factors. They are SINR function to measure probability factor when choosing corresponding actions and design of reward function assuming neighbor device and communication range. With these two functions, a latency enhanced with minimum data loss is achieved to attain better connectivity for D2D communication. At last, the neighbor device and communication range for corresponding data stream are employed to reduce data loss and latency for D2D communication. The simulation result illustrates that the BCO-DLQL method increases the energy efficiency by 7% and reduces the latency by 24% as compared to state-of-the-art works.

KEYWORDS:

• Bio-inspired
• Conic
• Fifth generation
• Distributed latency
• Particle swarm optimization
• Q learning

Additional information

Notes on contributors

Sridhar Varadala

V. Sridhar is working as assistant professor in ECE Department at Vidya Jyothi Institute of Technology, Hyderabad, Telangana, India. Presently pursuing (PhD) at Sathyabhama Institute of Science and Technology, Telangana, India, completed MTech with specialization in wireless and mobile communication from JNTUH. Completed Electronics and telecommunication engineering from JNTUH. He has published 35 international journals and 4 scopus papers and one Web of Science paper. His areas of research interests include wireless and mobile communications, artificial intelligence, machine learning, optical fiber communication image processing, telecommunications, communication systems, signal processing. He is Lifetime Member of ISTE, IETE, IAENG, and SDIWC.

S. Emalda Roslin

S Emalda Roslin received her bachelor's degree in electronics and communication engineering from St Josephs College of Engineering, Chennai in 2000. She received master's degree in applied electronics from Sathyabama University, Chennai, Tamil Nadu, India, in 2004 and her PhD in 2013 from Sathyabama University. She completed her doctoral research in “An Energy Efficient Topology Control Algorithm for Enhanced Network lifetime in Wireless Sensor Networks”. She is currently working as a professor in the Department of Electronics and Communication Engineering, Sathyabama Chennai, Tamil Nadu, India and she is having 19 years of teaching experience. She has published more than 50 research papers in various international/national conferences and international/national journals indexed in Web of Science/ Scopus/other databases. She is guiding around seven research scholars. Her major areas of interest are wireless multimedia networks, wireless mesh networks and image processing. She is also an editorial board member for various journals like International Journal of Artificial Intelligence and Mechatronics, American Journal of Science, Engineering and Technology, International Journal of Advanced Smart Sensor Network Systems, etc. Email: [email protected]