China
Africa
Explore chapters and articles related to this topic
Toward Quantum-Inspired SSA for Solving Multiobjective Optimization Problems
Published in Siddhartha Bhattacharyya, Mario Köppen, Elizabeth Behrman, Ivan Cruz-Aceves, Hybrid Quantum Metaheuristics, 2022
Siddhartha Bhattacharyya, Mario Köppen, Elizabeth Behrman, Ivan Cruz-Aceves
The source code of MQSSA is developed in MATLAB R2017b, and the experimental computer environment for the simulation is used as follows: Intel® Core™ i7-3520M CPU @ 2.90 GHz16 GB of RAMmacOS Catalina v.10.15.7 operating system
Predicting Driver Fatigue in Monotonous Automated Driving with Explanation using GPBoost and SHAP
Published in International Journal of Human–Computer Interaction, 2022
Feng Zhou, Areen Alsaid, Mike Blommer, Reates Curry, Radhakrishnan Swaminathan, Dev Kochhar, Walter Talamonti, Louis Tijerina
The proposed model used both physiological and behavioral measures to predict driver fatigue indicated by the PERCLOS measure. The model used 10 predictor variables as input and the model with the best performance only made use of heart data and breathing data and was able to predict driver fatigue with high accuracy in real time. With Python 3.8 on a MacBook Pro with 2.3 GHz Quad-Core Intel Core i7 and macOS Catalina, the average prediction time for one sample was only seconds. Compared with the fatigue-related studies in automated driving, our work used prediction models rather than simply described the fatigue progression in automated driving. In addition, compared to many machine learning models with potential identity confounding in fatigue detection in manual driving (Khan & Mansoor, 2008; McDonald et al., 2014), our model not only took the within-subjects correlation into consideration enabled by GPBoost with great performance as shown in Table 2, but also applied SHAP to explain the main effects between the predictor variables and driver fatigue.
Utilisation of machine learning algorithms for the prediction of syngas composition from biomass bio-oil steam reforming
Published in International Journal of Sustainable Energy, 2021
Adewale George Adeniyi, Joshua O. Ighalo, Gonçalo Marques
The proposed methods have been designed and validated using the Orange Data Mining Software, a component-based data mining framework for data visualisation and analytics. This software is open-source software and supports GPL licence. The code is hosted online and available for user analysis. This software is developed in Python programming language and compatible with Linux, Mac OS and Windows operating systems (Demšar et al. 2013). The proposed models have been designed using the version 3.23.1 of the software on a MacBook Pro (v.2018) running on a macOS Catalina version 10.15.3 operating system. For the modelling procedure, the model was developed and tested using two different methods such as testing on training data and stratified 10-fold cross-validation. In this methodology, the software shuffles the data is shuffled into 10 folds, takes a fold as test data and considers the rest as training data. In the input layer, there are 3 nodes (featurevariables), and the output layer has each of the four target variables. Usually, the neural network learns from the dataset and then uses it to make predictions of the target variables at various levels of the input feature.
Optimal pricing policies for tandem queues: Asymptotic optimality
Published in IISE Transactions, 2020
Throughout this section, we study three-station tandem queueing systems (i.e., J = 3) with buffer sizes and various B1. We set and consider three types of distribution F: exponential, uniform, and normal. We apply the unichain policy iteration algorithm (Puterman, 1994, p. 378) to determine the optimal dynamic gain (i.e., the optimal objective value in problem (3)). For the optimal static gain (i.e., the optimal objective value in problem (5)), we evaluate the gain under each static pricing policy πa, where using Equation (8.2.1) in Puterman (1994) and choose the highest value as the optimal static gain. Meanwhile, the simple static gain can be calculated using obtained by solving problem (6). All the algorithms are implemented with Matlab R2017b under MacOS Catalina on a MacBook with a 3.5 GHz Dual-Core Intel i7 processor and 16 GB RAM.