ADAMS Lab's repositories
amr-samples-metamodels-package
This repository contains data for training and testing metamodels (Kriging, RBF,...) that are used in a new surrogate based optimization (SBO) process called adaptive model refinement. The resulting trained metamodels are also included in this package.
Getting-Set-Up
Stemming from a summer-long project in the ADAMS Lab, this guide is meticulously crafted to provide step-by-step instructions, tips, and troubleshooting methods to ensure seamless integration and use of Crazyflies for both newcomers and experienced users.
PEMF
Predictive Estimation of Model Fidelity (PEMF) is a generalized approach to predict surrogate model error. PEMF takes as input a model trainer, sample data on which to train the model, and hyper-parameter values to apply to the model. As output, it provides an estimate of the median or maximum error in the surrogate model.
UAV-Noise-Experimental-Data-and-Modeling
A data set of experimental recordings of the acoustic sound pressure level field around a hovering UAV, and a Sequential Physics Infused machine learning model to predict the acoustic field based on the data.
optimizing-bio-inspired-riblets-openfoam
This repository contains key software artifacts of the Open-FOAM based CFD framework that computes the drag coefficient of a NACA-0012 airfoil with bio-inspired surface riblets. This package also provides the framework to introduce your own optimization code for minimizing drag by varying the shape of the surface riblets. For further information about the methods underlying this software package, refer to our paper: Lulekar, S. S., Ghassemi, P., Alsalih, H., and Chowdhury, S., An Adaptive-Fidelity Design Automation Framework toExplore Bio-inspired Surface Riblets for Drag Reduction, AIAA Journal. 2020. AIAA. The following sections provide further information on the usage of this software package:
Hybrid-Genetic-Algorithm-with-Adaptive-Crossover-for-Path-Planning
A new adaptive crossover method is proposed to solve the Multi Travelling Salesman Problem (mTSP) using real parameter Genetic Algorithm, to minimize the longest tour. The crossover operator used is dynamically changed with respect to the spread of the distances travelled by each salesman. A two-part chromosome is used to represent each candidate solution and crossover method is designed for the same.