Innovative Soft-Computing Solutions for Industrial and Environmental Problems

Abstract

Novel solutions, based on soft-computing techniques, are proposed in the present issue. All of them target open problems in the environmental and industrial domains. Thanks to the intelligent systems that are presented, the addressed problems are solved in innovative ways, advancing the present solutions. Deep learning is proposed in the first paper for predicting energy consumption in the residential domain. The target is explaining the impact of the input attributes on the prediction by taking into account the long-term and short-term properties of the time-series forecasting. The model consists of several components: two encoders represent the power information for prediction and explanation, a decoder predicts the power demand from the concatenated outputs of encoders, and an explainer identifies the most significant attributes for predicting the energy consumption. Several experiments on a benchmark dataset of household electric energy demand show that the proposed method explains the prediction appropriately with the most influential input attributes in the long-term and short-term dependencies. There is a trade off between the gain of the time-series explanation of the result and the prediction performance (slightly degraded). The second contribution also addresses a challenge in the energy field, as a thermal solar generation system is studied. The performance of four clustering techniques, with the objective of achieving strong hybrid models in supervised learning tasks, are compared. A real dataset is studied to validate several cluster methods when subsequently applying a regression technique to predict the output temperature of the system. With the objective of defining the quality of each clustering method, two approaches have been followed. The first one is based on three unsupervised learning metrics (Silhouette, Calinski-Harabasz and Davies-Bouldin) while the second one employs the most common error measurements for a regression algorithm (the MultiLayer Perceptron). Basurto et al. predict, by Supervised Machine Learning, the success of Private Participation Projects in the Telecom sector. Widely acknowledged classifiers (k Nearest Neighbors, Support Vector Machines, and Random Forest) are applied to an open dataset from the World Bank. The results on this highly imbalanced dataset are greatly improved by the application of data balancing techniques. It includes some standard ones (Random Oversampling, Random Undersampling, and SMOTE), together with some other advanced ones (Density-Based SMOTE and Borderline SMOTE). The satisfactory results validate the proposed application of classifiers on the dataset improved by data-balancing techniques. Supply chain network design (SCND) is the process for designing and modeling the supply chain, trying to minimize the costs generated by the location of facilities and the flow of product between the selected facilities. The aim of the fourth contribution is to investigate a particular SCND, namely the two-stage supply chain network design problem with risk-pooling and lead times. To do so, a novel efficient and effective genetic algorithm, designed to fit the challenges of the considered optimization problem,