Reinforcement Learning in Adaptive Control of Power System Generation

Raju, Leo; Milton, Robert; Suresh, Swetha; Sankar, S.

doi:10.1016/j.procs.2015.02.012

Cited by 13 publications

(8 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Reinforcement learning is a type of machine learning which is focused on goal-directed learning from interaction [14]. RL is an efficient method for sequential decision-making problems, making them ideal for network management [15]. In RL, the learner agent takes actions, and each action receives a reward as a feedback signal.…”

Section: Background Of Reinforcement Modelmentioning

confidence: 99%

Base Station Power Optimization for Green Networks Using Reinforcement Learning

Aktaş

Alemdar

2021

Sakarya University Journal of Computer and Information Sciences

View full text Add to dashboard Cite

The next generation mobile networks have to provide high data rates, extremely low latency, and support high connection density. To meet these requirements, the number of base stations will have to increase and this increase will lead to an energy consumption issue. Therefore "green" approaches to the network operation will gain importance. Reducing the energy consumption of base stations is essential for going green and also it helps service providers to reduce operational expenses. However, achieving energy savings without degrading the quality of service is a huge challenge. In order to address this issue, we propose a machine learning based intelligent solution that also incorporates a network simulator. We develop a reinforcement-based learning model by using deep deterministic policy gradient algorithm. Our model update frequently the policy of network switches in a way that, packet be forwarded to base stations with an optimized power level. The policies taken by the network controller are evaluated with a network simulator to ensure the energy consumption reduction and quality of service balance. The reinforcement learning model allows us to constantly learn and adapt to the changing situations in the dynamic network environment, hence having a more robust and realistic intelligent network management policy set. Our results demonstrate that energy efficiency can be enhanced by 32% and 67% in dense and sparse scenarios, respectively.

show abstract

Section: Background Of Reinforcement Modelmentioning

confidence: 99%

Base Station Power Optimization for Green Networks Using Reinforcement Learning

Aktaş

Alemdar

2021

Sakarya University Journal of Computer and Information Sciences

View full text Add to dashboard Cite

show abstract

“…UC focuses on scheduling generating units to meet the forecasted load demand over a time period under different operational constraints (Coronado et al, 2012). Its prime objective is to reduce total generation cost (Raju et al, 2015). Santillan et al (2016) were able to show that failure to implement UC model in a peaking power plant increases generation costs at approximately 27%.…”

Section: Unit Commitmentmentioning

confidence: 99%

Policy-based unit commitment problem: a case in a peaking power generation plant

Jr.

Santillan

Senen

et al. 2017

IJADS

View full text Add to dashboard Cite

Unit commitment (UC) along with economic load dispatch (ELD) is an integral optimisation problem in the power generation industry. In this paper, policy-based algorithms are developed to address dispatch problems under different system configurations with the benefit of providing flexibility in addressing several desired policies of decision-makers. A case study is conducted in a diesel-fired, power plant in central Philippines to elucidate the proposed approach. Three policies were created to effectively address the optimisation problem: 1) classic UC model; 2) buying and selling strategy; 3) continuous loading strategy with various issues being considered. Results show that the strategies exhibit a significant decrease in total costs and ease of implementation with respect to the current production schedule of the case firm. Furthermore, the second policy provides the least generated cost which implies that generating firms must consider the option to sell power when market prices reach desired levels.

show abstract

“…Unit Commitment. Raju et al [1] described a unit commitment problem (UC) as an elementary level scheduling process of power systems. While economic load dispatch (ELD) assumes that all generating units are online, UC determines which units to dispatch, start up, shut down, ramp up, or ramp down.…”

Section: Literature Reviewmentioning

confidence: 99%

“…UC focuses on scheduling generating units to meet the forecasted load demand over a time period or planning horizon under different operational constraints [9]. Its prime objective is to reduce total generation costs particularly fuel costs [1]. With the UC problem, ELD becomes a subproblem.…”

Section: Literature Reviewmentioning

confidence: 99%

See 1 more Smart Citation

Application of Unit Commitment with Market Pool in a Peaking Power Generation Firm in the Philippines for a Cost Reduction Case Study

Santillan

Senen

Wamilda

et al. 2016

Journal of Engineering

View full text Add to dashboard Cite

This paper attempts to provide an optimum loading schedule of power generating units with the least cost by solving a unit commitment (UC) problem and to present good estimates of cost differences when UC problem is not applied. UC is a fundamental optimization problem of power generation systems which determines the optimum schedule of generating units which minimizes generation costs. However, for small power generation firms which are situated in developing countries, UC-based problems are poorly understood if not implemented and the scheduling of generating units is based on some methodologies which may provide results that are not optimal. Thus, a case study in a small power generation firm in central Philippines is carried out to elucidate these objectives. The case requires a solution of the mixed-integer nonlinear programming (MINLP) problem. Results show that the proposed UC-based problem yields optimal costs and the cost disparity from the current scheduling scheme is approximately at 27% which may be considered as potential cost savings. This shows that UC-based problem provides a reliable platform in achieving minimal generation costs. These results are significant to decision-makers particularly in small power generation firms and to engineering practitioners in the Philippines and in some developing countries as these provide an overview of the disparity of cost figures of not implementing UC.

show abstract

Reinforcement Learning in Adaptive Control of Power System Generation

Cited by 13 publications

References 10 publications

Base Station Power Optimization for Green Networks Using Reinforcement Learning

Base Station Power Optimization for Green Networks Using Reinforcement Learning

Policy-based unit commitment problem: a case in a peaking power generation plant

Application of Unit Commitment with Market Pool in a Peaking Power Generation Firm in the Philippines for a Cost Reduction Case Study

Contact Info

Product

Resources

About