In Cellular Network Base Stations data utilization depend on various factors. Data utilization patterns by using Machine Learning (ML) algorithms can be studied. Multiple servers are operational in Base Stations at full capacity to ensure 24 × 7 services however the actual utilization of resources may be less as per the demand for data services in the area. Hence, the server hardware resource allocation can be controlled (linked) as per requirement during the day. To achieve this historical cellular traffic data can be used for identifying trends and patterns which is based on activities in the area. In this chapter, the aim is to design a resource scheduling method based on machine learning and the same can be used for preserving power consumption in cellular base stations. An experimental study between popular ML approaches is done comparing the performance of unsupervised & supervised learning algorithms. Further Long Sort Term Memory (LSTM) & ANN model training is done for predicting future workload for the next 24 h & 7 days. Thus, a scheduling algorithm has been proposed based on a predicted 24-hour workload, and a 7-day workload. This will switch OFF/ON the server hardware processing units in Base Stations based on the requirement, otherwise, in a normal scenario, these server units remain idle and consume power. The experimental data is refereed from the Kaggle repository for 4G (LTE Traffic Prediction). As per the theoretical and basic experimental analysis, we can say that the proposed technique will be able to reduce approx. 25% energy consumption by temporarily switching off the server devices.

Authors and Affiliations

1. Department of CSE, SAGE University, Indore, M.P, India

   Prashant Shrivastava & Sachin Patel

Corresponding author

Correspondence toPrashant Shrivastava.

Editors and Affiliations

1. Bharath Institute of Higher Education and Research, Chennai, India

   Rabindra Nath Shaw

2. Systems Research Institute, Warsaw, Poland

   Marcin Paprzycki

3. The Neotia University, Sarisha, India

   Ankush Ghosh

Reprints and permissions

© 2023 The Author(s), under exclusive license to Springer Nature Switzerland AG

Policies and ethics