High Impact Factor : 4.396 icon | Submit Manuscript Online icon |

Custom Application in Salesforce using Augmented Random Search

Author(s):

Prof. Kanchan Jadhav , Sinhgad Academy of Engineering; Kalpraj Dakhore, Sinhgad Academy of Engineering; Rushabh Gupta, Sinhgad Academy of Engineering; Mayur Khandare, Sinhgad Academy of Engineering; Sahil Mahale, Sinhgad Academy of Engineering

Keywords:

Augmented Random Search, Salesforce

Abstract

India is a developing country and in various developing The increasing demands of Robots and Robotic Simulation all over the world has given a new dimension to the Automation and Robotics Industry. The physical training and repairing of a robot in real environment can adversely affect the costs, thereby reducing the efficiency. Reinforcement Learning has proved out to be most prominent way to overcome this issue with the help of Simulations and Agent based modelling. The Reinforcement learning has Basic Random Search algorithm which is used to perform such competitive tasks but it limitates due to high variance in random values generated. In this paper, we are implementing a model free reinforcement learning algorithm known as Augmented Random Search Algorithm which uses Shallow Learning Neural network and Method of Finite Differences. The simplified policies and derivative free methods make this algorithm simple to work. ARS is at least 15X faster than Basic Random Search Algorithm, but it can be made more efficient with the use of cloud rather than using local system hardware. That’s where the use of Salesforce helps us to utilize the cloud environment.

Other Details

Paper ID: IJSRDV7I100253
Published in: Volume : 7, Issue : 10
Publication Date: 01/01/2020
Page(s): 399-401

Article Preview

Download Article