DataFITR: An Open, Guided Input Modeling Tool for Creating Simulation-Based Digital Twins

Lekshmi P., Tushar Lone, Neha Karanjkar

2023

Abstract

Input Modeling (IM) is a critical step in the process of building simulation-based digital twins. It involves selecting a family of distributions to model the observed data and finding the distribution parameter values that best fit the data. Subsequently, random variates adhering to the selected distribution can be generated to create a simulation-based digital twin of the system. For complex systems, IM can be a nuanced process involving a series of decisions that require visual feedback at each step. There is currently a dearth of open, GUI-based tools for aiding the non-expert user in the process of IM. This paper presents DataFITR, a GUI-based, open Input Modeling tool we have developed for guiding the non-expert user through the steps of input modeling and automating several intermediate tasks. DataFITR is cloud-hosted with a web-based user interface. The user can upload data as a file and the tool guides the user through the IM process by suggesting types and suitable distributions for each observed variable. It generates multiple goodness-of-fit measures for a large set of standard discrete and continuous distributions and can also support arbitrary (non-standard) distributions using a Kernel Density Estimation approach. DataFITR also assists in exploratory data analysis by providing various statistical properties of the observed data and in finding correlations between output measures. Once a matching distribution is found, the tool generates Python code for producing random variates from the matching distribution, which can be directly inserted into a simulation model. In this paper, we describe the DataFITR tool and its features, and compare it with existing open libraries and tools for assisting IM. We present a simulation case study of a bottling plant to demonstrate the utility of the DataFITR tool in building simulation-based digital twins.

Paper Citation

in Harvard Style

P. L., Lone T. and Karanjkar N. (2023). DataFITR: An Open, Guided Input Modeling Tool for Creating Simulation-Based Digital Twins. In Proceedings of the 13th International Conference on Simulation and Modeling Methodologies, Technologies and Applications - Volume 1: SIMULTECH; ISBN 978-989-758-668-2, SciTePress, pages 279-286. DOI: 10.5220/0012082600003546

in Bibtex Style

@conference{simultech23,
author={Lekshmi P. and Tushar Lone and Neha Karanjkar},
title={DataFITR: An Open, Guided Input Modeling Tool for Creating Simulation-Based Digital Twins},
booktitle={Proceedings of the 13th International Conference on Simulation and Modeling Methodologies, Technologies and Applications - Volume 1: SIMULTECH},
year={2023},
pages={279-286},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0012082600003546},
isbn={978-989-758-668-2},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 13th International Conference on Simulation and Modeling Methodologies, Technologies and Applications - Volume 1: SIMULTECH
TI - DataFITR: An Open, Guided Input Modeling Tool for Creating Simulation-Based Digital Twins
SN - 978-989-758-668-2
AU - P. L.
AU - Lone T.
AU - Karanjkar N.
PY - 2023
SP - 279
EP - 286
DO - 10.5220/0012082600003546
PB - SciTePress