An intelligent demand forecasting model using a hybrid of metaheuristic optimization and deep learning algorithm for predicting concrete block production

Abstract

Demand forecasting aims to optimize the production planning of industrial companies by ensuring that the production planning meets the future demand. Demand forecasting utilizes historical data as an input to predict future trends of the demand. In this paper, a new approach for developing an intelligent demand forecasting model using a hybrid of metaheuristic optimization and deep learning algorithm is presented. Firefly algorithmbased gated recurrent units (FA-GRU) is used to tackle the production forecasting problem. The proposed model has been evaluated and compared with the standard gated recurrent unit (GRU) and standard long short-term memory model (LSTM) using historical data of 36 months of concrete block manufacturing at dler company in Iraq. The prediction accuracy of the three models is evaluated using the root mean square error (RMSE), the mean absolute percentage error (MAPE) and the statistical coefficient of determination (R2 ) indicators. The outcomes of the study show that the proposed FA-GRU gives better forecasting results compared to the standard GRU and standard LSTM.

A Systematic Review of the Literature on the Use of Artificial Intelligence in Forecasting the Demand for Products and Services in Various Sectors

“…Finally, we have that this group is made up of 13 documents belonging mainly to the energy and retail sectors, whose main characteristic is the use of deep learning in their forecast models. The models proposed by this group are aimed at improving the performance of deep learning algorithms through various resources, among which are: the use of hyper-parameter optimization algorithms, such as the "firefly algorithm" [31] or the " Improved Giza pyramids construction algorithm" [32]; the use of dimensionality reduction techniques such as "Encoders" [33], [34] and "Principal Component Analysis" (PCA) to optimize the model inputs; the use of "cross or transfer learning", that is, the use of data from similar products or services when the data of the product under study is very limited [35], [36], [37]; the use of "clustering" to divide the data into groups of similar behavior and train a neural network for each cluster [38]; the transformation of the data into images and their decomposition to then use a CNN for feature extraction and an LSTM for prediction [39]; the use of complex time series decomposition algorithms using neural networks [40]; the use of parallel computing [41]; and the use of special architectures of convolutional networks [42].…”

Dissecting the compensation conundrum: a machine learning-based prognostication of key determinants in a complex labor market

“…This study uses techniques such as accuracy, missing value imputation, outlier detection, normalization, autocorrelation among variables, cross-validation, out-of-sample testing and comparative analysis of model outcomes with alternative models or benchmarks to undergird the robustness checks. To determine the best-performing model, the root mean squared error (RMSE) metric, a reliable measure of accuracy (Karunasingha, 2022;Al-Khazraji et al, 2022), and correlation coefficient parameters are utilized. The CPM is approached as a multivariate regression problem where the output variable is compensation.…”

Optimization of Sliding Mode and Back-Stepping Controllers for AMB Systems Using Gorilla Troops Algorithm

“…Meta-heuristic algorithms are one of the well-known and fast growing optimization methods inspired by the behaviors observed in nature. These algorithms have been successfully used to solve a wide range of optimization problems [11][12][13][14]. Motivated by powerful of these algorithm and to get the optimal behavior of the proposed controller, this paper is assigned the tuning process to the gorilla troops optimization (GTO) to tune the adjustable parameters of the SMC and BSC controllers.…”