Accurate Demand Forecasting Research Articles

Comparative study of univariate and multivariate strategy for short-term forecasting of heat demand density: Exploring single and hybrid deep learning models

Accurate short-term forecasting of heating energy demand is needed for achieving optimal building energy management, cost savings, environmental sustainability, and responsible energy consumption. Furthermore, short-term heating energy prediction contributes to zero-energy building performance in cold climates. Given the critical importance of short-term forecasting in heating energy management, this study evaluated six prevalent deep-learning algorithms to predict energy load, including single and hybrid models. The overall best-performing predictors were hybrid models using Convolutional Neural Networks, regardless of whether they were multivariate or univariate. Nevertheless, while the multivariate models performed better in the first hour, the univariate models often were more accurate in the final 24 h. Thus, the best-performing predictor of the first timestep was a multivariate hybrid Convolutional Neural Network–Recurrent Neural Network model with a coefficient of determination (R²) of 0.98 and the lowest mean absolute error. Yet, the best-performing predictor of the final timestep was the univariate hybrid model Convolutional Neural Network–Long Short-Term Memory with an R² of 0.80. Also, the prediction accuracy of the best-performing multivariate hybrid models reduced faster per hour compared to the univariate models. These findings suggest that multivariate models may be better suited for early timestep predictions, while univariate models may be better suited for later time steps. Hence, combining the models can enhance accuracy at various timesteps for achieving high fidelity in forecasting and offering a comprehensive tool for energy management.

New energy power demand prediction and optimal scheduling based on artificial intelligence in smart grid

Abstract With the development of smart grid, the demand for new energy power increases. Improving the accuracy of new energy power demand forecast is an important basis for the orderly operation of power system. This article presents a new energy power demand forecasting method based on DESSA-NESN algorithm. First, differential evolution algorithm (DE) and sparrow search algorithm (SSA) are combined, and operations such as mutation, crossing and screening are introduced into the population updating process of SSA. The internal state function of the savings pool of the standard echo state network (ESN) is replaced by the hyperbolic tangent function to obtain the nonlinear echo state network (NESN). Then, the parameters of deep echo state network (DESN) are optimized using DESSA algorithm. The DESSA-DESN prediction model is established. Finally, the mean absolute percentage error (MAPE) and root mean square error (RMSE) of DESSA-NESN were 15.84 and 0.12%, respectively, and the prediction effect was better than other comparison models.

A Study on Artificial Intelligence – Driven Inventory Solutions and Enhancing Manufacturing Efficiency

Artificial Intelligence is pivotal in transforming Manufacturing and Inventory management practices and maintaining this delicate inventory balance. AI-driven predictive forecasting, for instance, considers historical data, seasonality, and market dynamics to provide exceptionally accurate demand forecasts. Managing inventory has become a complex puzzle in today’s fast-paced business world. As companies expand their reach and cater to a global audience, tracking inventory and maintaining an accurate “available-to-promise” status in real time has become a challenge. In the current competitive landscape, retailers prioritise a seamless approach to inventory management, especially with the growing importance of omnichannel fulfilment. The art of balancing supply and demand while keeping costs in check has become a high-stakes game where every piece of the inventory puzzle must fit perfectly. In this article, we explore the remarkable role of Artificial intelligence in Manufacturing and Inventory management. We will discuss how businesses can harness the power of AI to optimize their efficiency in Manufacturing and Inventory processes.

Forecasting High Speed Diesel Demand in India with Econometric and Machine Learning Methods

According to International Energy Agency (IEA), India is expected to surpass China by 2024 to become the second largest consumer of oil in the world followed by the United States. High-Speed Diesel (HSD) has the biggest share in the total petroleum products consumed in India accounting for around 38% of the total consumption. Considering the volatile global oil market and an oil import dependency ratio of more than 80% during the last four years, the probability of supply disruptions is high in the Indian context. As any uncertainty about the supply of diesel can affect the smooth functioning of the economy and may create inflationary pressures. Accurate forecasting of HSD demand will be essential for appropriate supply management arrangements. Artificial Neural Networks (ANN) with Multi-Layer Perceptron (MLP) and extreme learning machines is used for forecasting diesel demand in this study. Demand forecasting has been carried out using monthly HSD demand data drawn from the “Indiastat” database for the period 1991-2022. Comparison of ANN with traditional forecasting methods of Autoregressive Integrated Moving Average(ARIMA)and Exponential Smoothing has also been undertaken in this study. This study identifies the deep learning technique of ANN with MLP as the best diesel demand forecasting technique.

An integrated approach for modern supply chain management: Utilizing advanced machine learning models for sentiment analysis, demand forecasting, and probabilistic price prediction

In the contemporary business landscape, effective interpretation of customer sentiment, accurate demand forecasting, and precise price prediction are pivotal in making strategic decisions and efficiently allocating resources. Harnessing the vast array of data available from social media and online platforms, this paper presents an integrative approach employing machine learning, deep learning, and probabilistic models. Our methodology leverages the BERT transformer model for customer sentiment analysis, the Gated Recurrent Unit (GRU) model for demand forecasting, and the Bayesian Network for price prediction. These state-of-the-art techniques are adept at managing large-scale, high-dimensional data and uncovering hidden patterns, surpassing traditional statistical methods in performance. By bridging these diverse models, we aim to furnish businesses with a comprehensive understanding of their customer base and market dynamics, thus equipping them with insights to make informed decisions, optimize pricing strategies, and manage supply chain uncertainties effectively. The results demonstrate the strengths and areas for improvement of each model, ultimately presenting a robust and holistic approach to tackling the complex challenges of modern supply chain management.

РОЛЬ ИНФОРМАЦИОННЫХ СИСТЕМ В ОПТИМИЗАЦИИ ЛОГИСТИКИ И ЦЕПОЧЕК ПОСТАВОК

With accelerated globalisation and ever-changing market conditions, effective logistics and supply chain management is becoming a critical success factor for companies of all sizes. Information systems enable more accurate demand forecasting, inventory optimisation and shorter delivery times, which ultimately contribute to the competitiveness of enterprises. An important component of effective logistics supply chain management is the integration of digital technologies. The research conducted analyses five key technologies – CRM, SCM, APS, MES and MRP, identifying their advantages and disadvantages. The impact of artificial intelligence in optimising logistics processes is also explored. The article also examines the process of global digitalisation in the logistics sector, including the adoption of the latest technologies and an analysis of possible threats to data security. Successful examples of the use of these technologies in various logistics companies are highlighted, and recommendations are given on the development of information technology infrastructure of enterprises and support for innovative projects in the field of logistics for effective adaptation to the rapidly changing digital requirements of society.

Exploring Heat Demand Forecasting in District Heating Networks Using Random Parameter Linear Regression Model

Abstract Accurate forecasting of heat demand in district heating networks is essential for their efficient and sustainable operation. This paper presents a novel approach using a random parameter linear regression model to forecast heat demand, distinguishing itself from classical linear regression models by its ability to address unobserved heterogeneity among parameters. Through a case study in Estonia and utilizing data from 2018 to 2023 and considering seasonality and consumption patterns, the study investigates determinants of heating demand in district heating networks. Two models were trained for heating and non-heating seasons. Results indicate significant impacts of weather conditions, energy prices, time of day, and network infrastructure on heat supply during the heating season, while only time of day and electricity prices were significant drivers during the non-heating season, with no notable influence of weather conditions. Prediction accuracy was slightly enhanced using the random parameter linear regression model, with a mean absolute percentage error of 9.66 % compared to 9.99 % for the Multi Linear Regression Model on the testing set.

Forecasting air passenger traffic and market share using deep neural networks with multiple inputs and outputs.

In this study, we address the challenge of accurate time series forecasting of air passenger demand using historical market demand data from the U.S. commercial aviation industry in the 21st century. Commercial aviation is a major contributor to the U.S. economy, directly or indirectly generating ~US$1.37 trillion annually, or 5% of annual GDP, and supporting more than 10 million jobs (Airlines for America, 2024). Over 1 billion passengers flew through U.S. airports in 2023 (Bureau of Transportation Statistics, 2024a). Using multiple correlated time series inputs predicts future values of multiple interrelated time series and leverages their mutual dependencies to enhance accuracy. In this study, we introduce a two-stage algorithm employing a deep neural network for correlated time series forecasting, addressing scenarios where multiple input variables are interrelated. This approach is designed to capture the influence that one time series can exert on another, thereby enhancing prediction accuracy by leveraging these interdependencies. In the first stage, we fit four Recurrent Neural Network (RNN) models to generate accurate univariate forecasts, each functioning as a single input-output model to predict aggregated market demand. The Gated Recurrent Unit (GRU) model was the top performer for our dataset overall. In the second stage, we apply the best fitted model (GRU Model) from Stage 1 to each individual competitor (disaggregated from the market) and then merge all input tensors using the Concatenate function. We hope to contribute to the relevant body of knowledge with a deep neural network framework for forecasting market share among competitors in the U.S. commercial aviation industry, as no similar approach has been documented in the literature. Given the importance of the industry, there is potentially great value in applying sophisticated forecasting techniques to achieve accurate predictions of air passenger demand. Moreover, these techniques may have wider applications and can potentially be employed in other contexts.

Outliers in intermittent demand forecasting

The aim of the article was to analyze the impact of outliers on the accuracy of intermittent demand forecasting. The hypothesis was verified that eliminating the influence of outliers leads to more accurate forecasts. A data set containing monthly sales time series of over 12 000 products was analyzed. Three outliers rejection strategies were analyzed (replacing outliers, replacing extreme values, isolation forest). Intermittent demand forecasts were determined using the Croston, SBA, TSB and SES methods, with optimization of smoothing constants and initial values. The accuracy of the forecasts was assessed from the point of view of the inventories value and the value of lost sales (lost demand). The conducted research shows that eliminating the influence of outliers significantly reduces the inventory level value, with only a slight increase in the value of lost sales. Eliminating the influence of outliers therefore provides an opportunity to improve the financial results of enterprises selling slow – moving products through a significant reduction in inventory levels.

Demand Forecasting for Multi-Variety and Small-Batch Materials Based on Attention to Degree

Abstract Optimal production planning based on accurate market demand forecasting is crucial for cost control, inventory management, and market response in the multi-variety and small-batch production modes. Considering factors such as market demand, unit selling price, market demand trends, and product saturation within the attention cycle, an attention degree model for multi-variety and small-batch materials is constructed using historical market demand data from an electronic product manufacturing enterprise. A method for prioritizing and screening small-batch materials based on attention degree to formulate production plans is proposed. The demand for prioritized small-batch materials is predicted using the autoregressive integrated moving average model, multilayer perceptron, and bidirectional long short-term memory network. The optimal prediction results are selected to calculate the attention degree, which is then used to formulate the production plan for the next attention cycle to achieve orderly production. Taking the electronic product manufacturing enterprise as an example, the effectiveness and feasibility of the proposed model and method are verified by applying the prioritized production of small-batch materials screened based on attention degree.

Enhancing medium-term capacity management: Practical insights and considerations for decision support systems

Succeeding in volatile business environments requires agile operating companies. An effective approach to improve agility involves careful alignment of production capacity with customer demand. Medium-term capacity management, also known as aggregated planning, addresses this aspect. Today, operations managers are mainly relying on their experience when making these management decisions. Though, recent enhancements in machine learning, but also in simulation modeling and digital twins enable a further step towards the development of more advanced decision support systems. Such systems combining a digital image of the production system alongside accurate demand forecasts, utilizing advanced machine learning algorithms such as deep reinforcement learning, might increase the speed and quality of making these decisions. In this paper, we investigate practical considerations as prerequisites to build such a decision support system. We conducted an empirical survey involving 283 Austrian manufacturing companies to gain insight into their current needs and challenges in medium-term capacity management. Our investigation explores the main production capacity adjustment levers used in the companies. We further analyze how these survey findings can be translated into the design of a decision support system.

Capacity Strategies a Comparative Perspective in Manufacturing vs Service Industries

The main aim of this research is to provide a deeper understanding of how wise capacity planning can become a key success factor across various industrial sectors. It can have a positive impact on organizations by achieving optimal operational efficiency and delivering services or products that satisfy consumers. This study employs a qualitative approach focusing on literature review to explore and comprehend capacity strategies in a comparative context between manufacturing and service industries. Capacity planning holds significant importance for the expansion of all companies, particularly those operating in industrial sectors. Businesses can leverage these strategies to meet client needs, save on operational costs, and facilitate sustainable expansion. It’s crucial for companies to comprehend and address current obstacles. Key factors include accurate demand forecasting, effective company constrain management, and the ability to adapt and respond to market fluctuations. In today’s era of digitalization, companies have opportunities to employ sophisticated technologies like ERP, big data analytics, and machine learning to effectively achieve their capacity planning goals. Implementing these tools can result in more accurate forecasting and enable faster adaptation to market fluctuations. To enhance operational efficiency and align output with business objectives, companies need to engage in effective planning.

A Deep Learning Approach to Real-time Electricity Load Forecasting

In light of the increasing importance of accurate and real-time electrical demand forecasting, this research presents a deep learning model with the goal of dramatically improving predictive accuracy. Conventional methods of forecasting, such as linear regression, have trouble capturing the complex patterns included in data about electricity usage. Standard machine learning methods are shown to be wanting when compared to the suggested deep Long Short-Term Memory (LSTM) model. Mean Absolute Error (MAE) of 5.454 and Mean Squared Error (MSE) of 18.243 demonstrate the deep LSTM model's proficiency in tackling this problem. The linear regression, on the other hand, achieved a MAE of 47.352 and an MSE of 65.606, which is lower than the proposed model. Because of its greater predictive precision and reliability, the deep LSTM model is a viable option for accurate, real-time prediction of electricity demand.

An Improved Hybrid Approach for Daily Electricity Peak Demand Forecasting during Disrupted Situations: A Case Study of COVID-19 Impact in Thailand

Accurate electricity demand forecasting is essential for global energy security, reducing costs, ensuring grid stability, and informing decision making in the energy sector. Disruptions often lead to unpredictable demand shifts, posing greater challenges for short-term load forecasting. Understanding electricity demand patterns during a pandemic offers insights into handling future disruptions. This study aims to develop an effective forecasting model for daily electricity peak demand, which is crucial for managing potential disruptions. This paper proposed a hybrid approach to address scenarios involving both government intervention and non-intervention, utilizing integration methods such as stepwise regression, similar day selection-based day type criterion, variational mode decomposition, empirical mode decomposition, fast Fourier transform, and neural networks with grid search optimization for the problem. The electricity peak load data in Thailand during the year of the COVID-19 situation is used as a case study to demonstrate the effectiveness of the approach. To enhance the flexibility and adaptability of the approach, the new criterion of separating datasets and the new criterion of similar day selection are proposed to perform one-day-ahead forecasting with rolling datasets. Computational analysis confirms the method’s effectiveness, adaptability, reduced input, and computational efficiency, rendering it a practical choice for daily electricity peak demand forecasting, especially in disrupted situations.

Proposed Improvement of Demand Forecast Accuracy using Six Sigma DMAIC in PT XYZ

Proposed Improvement of Demand Forecast Accuracy using Six Sigma DMAIC in PT XYZ

Forecasting local hospital bed demand for COVID-19 using on-request simulations

Accurate forecasting of hospital bed demand is crucial during infectious disease epidemics to avoid overwhelming healthcare facilities. To address this, we developed an intuitive online tool for individual hospitals to forecast COVID-19 bed demand. The tool utilizes local data, including incidence, vaccination, and bed occupancy data, at customizable geographical resolutions. Users can specify their hospital’s catchment area and adjust the initial number of COVID-19 occupied beds. We assessed the model’s performance by forecasting ICU bed occupancy for several university hospitals and regions in Germany. The model achieves optimal results when the selected catchment area aligns with the hospital’s local catchment. While expanding the catchment area reduces accuracy, it improves precision. However, forecasting performance diminishes during epidemic turning points. Incorporating variants of concern slightly decreases precision around turning points but does not significantly impact overall bed occupancy results. Our study highlights the significance of using local data for epidemic forecasts. Forecasts based on the hospital’s specific catchment area outperform those relying on national or state-level data, striking a better balance between accuracy and precision. These hospital-specific bed demand forecasts offer valuable insights for hospital planning, such as adjusting elective surgeries to create additional bed capacity promptly.

IoT Water Management and Distribution System for Smart City using Artificial Intelligence

Year after year, the accessible percentage of the world's freshwater resources decreases. According to a World Economic Forum assessment, increased water consumption will result in catastrophic global shortages during the next two decades. The majority of water loss occurs during distribution in pipes during shipment. A water distribution system" using the Internet of Things (IoT) combining "cloud and fog computing" is presented for water distribution and underground pipeline health monitoring to remove the loss. Consumer needs must be assessed to design an effective IoT-based water supply system. As a result, a deep learning method known as Long short-term memory (LSTM) is suggested to estimate customer water consumption. An integrated IoT "Water Distribution Network (WDN)" is being created by employing hydraulic engineering and more accurate demand forecasts. The WDN design will minimize transport losses while maintaining water quality for users. This will result in the creation of smart systems for distributing water. Finally, the suggested algorithms' performance is analyzed and compared. The results demonstrate that the water distribution system with intelligence can effectively monitor the network.

Adaptive tourism forecasting using hybrid artificial intelligence model: a case study of Xi’an international tourist arrivals

Accurate forecasting of tourist demand is important to both business practitioners and government policymakers. In the past decade of rapid development of deep learning, many artificial intelligence methods or deep learning models have been built to improve prediction accuracy. But data-driven end-to-end deep network models usually require large data sets to support. For tourism forecasting, the sample is insufficient and many models are difficult to apply. In this article, we propose a novel hybrid model GM-LSTM, which combines the advantages of gray models and neural networks to achieve self-adaptive prediction with small samples. Specifically, the overall trend of tourism demand is captured by a first-order gray model and the non-linear residual fluctuation is characterized using a long short-term memory (LSTM) network with a rolling mechanism. The model is validated through a case study of up to 38 years of data on annual international tourist arrivals in Xi’an, China. The proposed GM-LSTM model achieved a predicted MAPE value of 11.88%, outperforming other time series models. The results indicate that our proposed hybrid model is accurate and efficient.

Improved Demand Forecasting of a Retail Store Using a Hybrid Machine Learning Model

Accurate demand forecasting is a competitive advantage for all supply chain components, including retailers. Approaches like naïve, moving average, weighted average, and exponential smoothing are commonly used to forecast demand. However, these simple approaches may lead to higher inventory and lost sales costs when the trend in demand is non-linear. Additionally, price strongly influences demand, and we can’t neglect the impact of price on demand. Similarly, the demand for a stock keeping unit (SKU) depends on the price of the competitor for the same SKU and the price of the competitive SKU. We thus propose a demand prediction model that considers historical demand data and the SKU price to forecast the demand. Our approach uses different machine-learning regressor algorithms and identifies the best machine-learning algorithm for the SKU with the lowest forecasting error. We further extend the forecasting model by training a hybrid model from the best two regression algorithms individually for each SKU. Forecasting error minimisation is the driving criterion for our literature. We evaluated the approach on 1000 SKUs, and the result showed that the Random Forest is the best-performing regressor algorithm with the lowest mean absolute percentage error (MAPE) of 8%. Furthermore, the hybrid model resulted in a lower inventory and lost sales cost with a MAPE of 7.74%. Overall, our proposed hybrid demand forecasting model can help retailers make informed decisions about inventory management, leading to improved operational efficiency and profitability.

Forecasting the Tourist Arrival Volumes and Tourism Income with Combined ANN Architecture in the Post COVID-19 Period: The Case of Turkey

Accurate forecasting of tourism demand and income holds paramount importance for both the tourism industry and the national economy. This study aims to address several objectives: (1) specify the best forecasting model in the prediction of tourist arrival volumes and tourism income for Turkey; (2) assess the degree of impact exerted by various determinants on the tourism forecasts; (3) generate forecasts for tourist arrival volumes and tourism income using the most suitable models; and (4) examine potential scenarios illustrating the ramifications of the Russia-Ukraine war on tourist arrival volumes and tourism income. The forecasting models employed in this study encompass a comprehensive set of statistical methods, including ETS, ARIMA, TRAMO-SEATS, X13, X11, STL, Grey, and their combinations with ANN. In the ANN models, exogenous variables such as the global financial crisis, the Turkey-Russia warplane crash crisis, the COVID-19 pandemic, and USD/TRY exchange rates are incorporated. The results unveil the identification of five superior models: ETS, Grey, hybrid ETS-ANN, hybrid Grey-ANN, and hybrid ARIMA-ANN models, which exhibit the lowest MAPE and sMAPE values. Forecasts for the forthcoming quarters are examined under two scenarios: assuming the continuity or cessation of the Russia-Ukraine war. Comparative analysis of the relative effects of exogenous variables indicates that COVID-19 has the most substantial impact on tourist arrival volumes, and tourism income is primarily influenced by the USD/TRY exchange rate.