Demand forecasting has become the most crucial part for supporting supply chain decisions. However, accurate forecasting in time series demand forecasting, particularly within supply chain ...operations, is challenging because of short-term data features, such as limited volume, nonlinear datasets, and near history disturbances. As one of the most promising deep learning models, long short-term memory shows superior performance in extracting implicit patterns from datasets of various areas. Thus, a novel forecasting framework, attLSTM is constructed combining enhanced bidirectional LSTM (LSTM) and self-attention mechanism. The forecasting performance of attLSTM is verified by testing six randomly selected datasets and eight additional datasets with different volumes from a given database. The proposed attLSTM is compared with seasonal autoregressive integrated moving average, support vector machine, random forest, and LSTM through two commonly applied evaluation metrics and a specially designed newsvendor cost model. Extended experiments are conducted on four benchmark datasets from other fields. These analyses demonstrate that attLSTM shows comparable performance in supporting the supply chain demand forecasting and operational decisions. The proposed framework has robust generalization capability in univariate time series demand forecasting.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPCLJ, UPUK, ZAGLJ, ZRSKP
With the availability of data and the increasing capabilities of data processing tools, many businesses are leveraging historical sales and demand data to implement smart inventory management ...systems. Demand forecasting is the process of estimating the consumption of products or services for future time periods. It plays an important role in the field of inventory control and Supply Chain, since it enables production and supply planning and therefore can reduce delivery times and optimize Supply Chain decisions. This paper presents an extensive literature review about demand forecasting methods for time-series data. Based on analysis results and findings, a new demand forecasting tool for inventory control is proposed. First, a forecasting pipeline is designed to allow selecting the most accurate demand forecasting method. The validation of the proposed solution is executed on Stock&Buy case study, a growing online retail platform. For this reason, two new methods are proposed: (1) a hybrid method, Comb-TSB, is proposed for intermittent and lumpy demand patterns. Comb- TSB automatically selects the most accurate model among a set of methods. (2) a clustering-based approach (ClustAvg) is proposed to forecast demand for new products which have very few or no sales history data. The evaluation process showed that the proposed tool achieves good forecasting accuracy by making the most appropriate choice while defining the forecasting method to apply for each product selection.
Recently there has been a significant proliferation in the use of forecasting techniques, mainly due to the increased availability and power of computation systems and, in particular, to the usage of ...personal computers. This is also true for power network systems, where energy demand forecasting has been an important field in order to allow generation planning and adaptation. Apart from the quantitative progression, there has also been a change in the type of models proposed and used. In the `70s, the usage of non-linear techniques was generally not popular among scientists and engineers. However, in the last two decades they have become very important techniques in solving complex problems which would be very difficult to tackle otherwise. With the recent emergence of smart grids, new environments have appeared capable of integrating demand, generation, and storage. These employ intelligent and adaptive elements that require more advanced techniques for accurate and precise demand and generation forecasting in order to work optimally. This review discusses the most relevant studies on electric demand prediction over the last 40 years, and presents the different models used as well as the future trends. Additionally, it analyzes the latest studies on demand forecasting in the future environments that emerge from the usage of smart grids.
Load forecasting is the most fundamental application in Smart-Grid, which provides essential input to Demand Response, Topology Optimization and Abnormally Detection, facilitating the integration of ...intermittent clean energy sources. In this work, several regression tools are analyzed using larger datasets for urban area electrical load forecasting. The regression tools which are used are Random Forest Regressor, k-Nearest Neighbour Regressor and Linear Regressor. This work explores the use of regression tool for regional electric load forecasting by correlating lower distinctive categorical level (season, day of the week) and weather parameters. The regression analysis has been done on continuous time basis as well as vertical time axis approach. The vertical time approach is considering a sample time period (e.g seasonally and weekly) of data for four years and has been tested for the same time period for the consecutive year. This work has uniqueness in electrical demand forecasting using regression tools through vertical approach and it also considers the impact of meteorological parameters. This vertical approach uses less amount of data compare to continuous time-series as well as neural network techniques. A correlation study, where both the Pearson method and visual inspection, of the vertical approach depicts meaningful relation between pre-processing of data, test methods and results, for the regressors examined through Mean Absolute Percentage Error (MAPE). By examining the structure of various regressors they are compared for the lowest MAPE. Random Forest Regressor provides better short-term load prediction (30 min) and kNN offers relatively better long-term load prediction (24 h).
•Urban area electrical energy demand forecasting using regression techniques.•Regional electrical load forecasting with influence of meteorological parameters.•Vertical time approach provides meaningful data correlation.•Random Forest Regressor provides better short-term load prediction (30 min).•kNN offers relatively better long-term load prediction (24 h).
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPCLJ, UPUK, ZAGLJ, ZRSKP
Prediction of taxi service demand and supply is essential for improving customer experience and provider's profit. Recently, graph neural networks (GNNs), modeling city areas as nodes in a ...transportation graph, have been shown efficient for this application as they utilize both local node features and the graph structure in the prediction. Still, further improvement can be achieved by either simultaneously exploiting different types of nodes/edges in the graphs or enlarging the scale of the transportation graph. However, both alternatives are challenged by the scalability of GNNs. An immediate remedy to the scalability challenge is to decentralize the GNN operation. In return, as shown by our theoretical analysis and experimentation, this creates prohibitively excessive node-to-node communication. In this paper, we first propose a heterogeneous GNN-LSTM algorithm for improved taxi-level demand and supply forecasting utilizing several edge types in the graph. Then, to enable the large-scale application of this approach, we propose a semi-decentralized GNN inference approach that achieves scalability at minimized communication and computation overheads. This is achieved by utilizing multiple cloudlets; data centers with moderate computation and communication capabilities that can fit at cellular base stations. Extensive experiments over real data show the advantage of the proposed GNN-LSTM algorithm in improving prediction accuracy, and the ability of the proposed semi-decentralized GNN approach in reducing the overall inference time by about an order of magnitude compared to centralized and decentralized inference schemes.
The rapid development of big data and smart technology in the natural gas industry requires timely and accurate forecasting of natural gas consumption on different time horizons. In this work, we ...propose a robust hybrid hours-ahead gas consumption method by integrating Wavelet Transform, RNN-structured deep learning and Genetic Algorithm. The Wavelet Transform is used to reduce the complexity of the forecasting tasks by decomposing the original series of gas loads into several sub-components. The RNN-structured deep learning method is built up via combining a multi-layer Bi-LSTM model and a LSTM model. The multi-layer Bi-LSTM model can comprehensively capture the features in the sub-components and the LSTM model is used to forecast the future gas consumption based on these abstracted features. To enhance the performance of the RNN-structured deep learning model, Genetic Algorithm is employed to optimize the structure of each layer in the model. Besides, the dropout technology is applied in this work to overcome the potential problem of overfitting. In this case study, the effectiveness of the developed method is verified from multiple perspective, including graphical examination, mathematical errors analysis and model comparison, on different data sets.
•This work focuses on hourly natural gas demand forecasting.•A novel hybrid RNN-based demand forecasting method is developed.•The effectiveness of the forecasting method is tested.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPCLJ, UPUK, ZAGLJ, ZRSKP
We present a probabilistic forecasting framework based on convolutional neural network (CNN) for multiple related time series forecasting. The framework can be applied to estimate probability density ...under both parametric and non-parametric settings. More specifically, stacked residual blocks based on dilated causal convolutional nets are constructed to capture the temporal dependencies of the series. Combined with representation learning, our approach is able to learn complex patterns such as seasonality, holiday effects within and across series, and to leverage those patterns for more accurate forecasts, especially when historical data is sparse or unavailable. Extensive empirical studies are performed on several real-world datasets, including datasets from JD.com, China’s largest online retailer. The results show that our framework compares favorably to the state-of-the-art in both point and probabilistic forecasting.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPCLJ, UPUK, ZAGLJ, ZRSKP
Accurate forecasts of natural gas demand can be essential for utilities, energy traders, regulatory authorities, decision makers and others. The aim of this paper is to test the robustness of a novel ...hybrid computational intelligence model in day-ahead natural gas demand predictions. The proposed model combines the Wavelet Transform (WT), Genetic Algorithm (GA), Adaptive Neuro-Fuzzy Inference System (ANFIS) and Feed-Forward Neural Network (FFNN). The WT is used to decompose the original signal in a set of subseries and then a GA optimized ANFIS is employed to provide the forecast for each subseries. ANFIS output is fed into a FFNN to refine the initial forecast and upgrade the overall forecasting accuracy. The model is applied to all distribution points that compose the natural gas grid of a country, in contradiction to the majority of the literature that focuses on a limited number of distribution points. This approach enables the comparison of the model performance on different consumption patterns, providing also insights on the characteristics of large urban centers, small towns, industrial areas, power generation units, public transport filling stations and others.
•The paper focuses in natural gas demand forecasting.•A novel hybrid computational intelligence model is proposed.•A large number of natural gas distribution points is examined.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UL, UM, UPCLJ, UPUK, ZRSKP
The tourism sector, with its perishable nature of products, requires precise estimation of demand. To this effect, we propose a deep learning methodology, namely Bayesian Bidirectional Long ...Short-Term Memory (BBiLSTM) network. BiLSTM is a deep learning model, and Bayesian optimization is utilized to optimize the hyperparameters of this model. Five experiments using the tourism demand data of Singapore are conducted to ascertain the validity and benchmark the proposed BBiLSTM model. The experimental findings suggest that the BBiLSTM model outperforms other competing models like Long Short-Term Memory (LSTM) network, Support Vector Regression (SVR), Radial Basis Function Neural Network (RBFNN) and Autoregressive Distributed Lag Model (ADLM). The study contributes to tourism literature by proposing a superior deep-learning method for demand forecasting.
•A novel deep learning model is proposed for tourism demand forecasting.•Bayesian optimization is employed to optimize the hyperparameters.•The effectiveness of the proposed model is validated via robustness analysis with multiple experiments.•The effect of multi-lagged variables on model performance is studied.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPCLJ, UPUK, ZAGLJ, ZRSKP