Indonesian Journal of Electrical Engineering and Computer Science Vol. 25, No. 1, January 2022, pp. 481~487 ISSN: 2502-4752, DOI: 10.11591/ijeecs.v25.i1.pp481-487  481 Journal homepage: http://ijeecs.iaescore.com The general design of the automation for multiple fields using reinforcement learning algorithm Vijaya Kumar Reddy Radha 1 , Anantha N. Lakshmipathi 2 , Ravi Kumar Tirandasu 3 , Paruchuri Ravi Prakash 4 1 Department of IT, Lakireddy Bali Reddy College of Engineering (Autonomus), Mylavaram, India 2 Department of CSE, Malla Reddy Engineering College (Autonomus), Secunderabad, Telangana, India 3 Department of CSE, KoneruLakshmaiah Education Foundation, Vaddeswaram, India 4 Department of IT, Prasad V. Potluri Siddhartha-Institute of Technology, Vijayawada, India Article Info ABSTRACT Article history: Received May 25, 2021 Revised Oct 25, 2021 Accepted Nov 21, 2021 Reinforcement learning is considered as a machine learning technique that is anxious with software agents should behave in particular environment. Reinforcement learning (RL) is a division of deep learning concept that assists you to make best use of some part of the collective return. In this paper evolving reinforcement learning algorithms shows possible to learn a fresh and understable concept by using a graph representation and applying optimization methods from the auto machine learning society. In this observe, we stand for the loss function, it is used to optimize an agent’s parameter in excess of its knowledge, as an imputational graph, and use traditional evolution to develop a population of the imputational graphs over a set of uncomplicated guidance environments. These outcomes in gradually better RL algorithms and the exposed algorithms simplify to more multifaceted environments, even though with visual annotations. Keywords: AutoML Computational graphs Loss function Recurrent neural network Reinforcement learning This is an open access article under the CC BY-SA license. Corresponding Author: Vijaya Kumar Reddy Radha Department of IT, Lakireddy Bali Reddy College of Engineering (Autonomus) Mylavaram, A. P, India E-mail: Vijayakumarr285@gmail.com 1. INTRODUCTION A long-standing goal of research into reinforcement learning is to blueprint of general purpose learning algorithms that can resolve an extensive array of issues. A probable resolution would be to devise a meta-learning technique that could model novel reinforcement learning algorithms that simplify to an extensive multiplicity of jobs automatically. In current years, automated machine learning (AutoML) has exposed huge success in automate the model of machine learning mechanism, such as neural networks architectures and design bring up to date rules [1], [2]. These previous procedures were intended for supervised learning but in reinforcement learning, there is additional mechanism of the algorithm that could be potential targets for model automation and it is not for all time clear with the best model, update process would be to put together these mechanism. Previous hard works for the computerization reinforcement learning algorithm detection have concentrate first and foremost on design modernize rules. These procedures learn the reinforcement learning update process itself and normally represent bring up to date rule with a neural network such as an recurrent neural network (RNN) or convolutional neural network (CNN), which can be professionally optimized with gradient-based techniques [3], [4]. There is only some profit of such an illustration. This demonstration is communicative enough to describe existing algorithms but also novel, undiscovered algorithms and also interpretable. This graph