Demand Forecast Error Research Articles

Analysis of the Impact of Different Forecasting Techniques on the Inventory Bullwhip Effect in Two Parallel Supply Chains with a Competition Effect

This paper studies the impact of different forecasting techniques on the inventory bullwhip effect in two parallel supply chains with the competition effect, which is in contrast to the situation of a single product in a serial supply chain. In particular, this paper constructs two parallel supply chains, each of which includes one manufacturer and one retailer. Moreover, the market demand is impacted by the self-price sensitivity coefficient, the cross-price sensitivity coefficient, the market share, and the demand shock. We then assumed that the retailer can forecast market demand by using different forecasting techniques (i.e., the moving average technique (MA), the exponential smoothing technique (ES), and the minimum mean square error technique (MMSE)). We constructed the quantity model of the bullwhip effect and the inventory bullwhip effect. Finally, we analyzed the impact of different forecasting techniques and market share on the inventory bullwhip effect. We analyzed the conditions under which the retailers should choose different types of forecasting techniques on the basis of the inventory bullwhip effect. The results show that the MMSE forecasting technique can reduce the lead-time demand forecast error to the largest extent, and the inventory bullwhip effect can obtain the lowest level using the MMSE method: retailer-1 can reduce the inventory bullwhip effect by using the MA technique, when the self-price sensitivity coefficient, the price autoregressive coefficient, and the probabilities associated with customers choosing retailer-1’s product are very low.

Chance-Constrained Water Pumping to Manage Water and Power Demand Uncertainty in Distribution Networks

Water pumping in drinking water distribution networks (WDNs) can be treated as a flexible load in the power distribution network (PDN). In this article, we formulate an optimization problem to minimize the electricity costs associated with pumping subject to WDN and PDN constraints. In practice, both water and power demands are uncertain and pumps should be scheduled to ensure that pump operation does not violate either networks’ constraints for nearly all possible uncertainty realizations. To address this problem, we formulate a chance-constrained (CC) optimization problem that simultaneously determines pumping schedules along with the parameters of real-time control policies that can be used to respond to water and power demand forecast errors. We use approximations and relaxations along with the scenario approach for CC programming to reformulate the optimization problem into a convex deterministic problem. We demonstrate the performance of the approach through case studies and also explore the impact of the relaxations, an approach to improve computational tractability, and tradeoffs associated with the way in which we define the cost of real-time control actions. We find that optimal scheduling and real-time control of water pumping can effectively manage water and power demand uncertainty, meaning water demand is satisfied and both the WDN and PDN operate within their limits; however, the approach is conservative leading to high reliability at high cost.

Analysis of the Fundamental Predictability of Prices in the British Balancing Market

This research analyses the non-linear and complex effects of drivers of system imbalance prices in the GB electricity market. Unlike day-ahead prices, the balancing settlement prices are comparatively under-researched, yet their importance is growing with greater market risks. The fundamental drivers of these prices are analysed over 2016-2019. The result of a non-linear modelling approach reveals that system imbalance price exhibits a regime-switching behaviour, driven by weather and demand forecast errors, as well as other system effects. Surprisingly, balancing prices are predictable out of sample and a regime switching specification is more accurate than a linear model for prediction.

Optimal control strategy of battery-integrated energy system considering load demand uncertainty

Uncertain load behavior affects the energy system operation scheduling in battery-integrated energy systems (BIES) due to the direct relation between optimal battery scheduling and load demand in the forthcoming time interval. In this paper, the optimal schedule of BIES is investigated with considering load demand uncertainty. The previous works have introduced a cost function and found optimal strategies using evolutionary optimization algorithms, which may fall in local minima. Against the previous works and in order to avoid falling in local minima, a closed-form expression is presented in this paper by analyzing a mathematically-modeled forecast error and considering battery characteristics. In the proposed method, the optimal battery charging/discharging state is analytically obtained based on forecast error and system parameters without the need to use numerical approximation algorithms. The applicability of the proposed scheduling strategy in the real energy system is shown by applying it to online optimal control of the battery in a real system with load demand forecast error, and also by comparing the results with previous methods. The proposed approach leads to the globally-optimal solution for battery scheduling and also shrinks the computational complexity of the BIES scheduling problem.

An Unintended Consequence of Platform Dependence: Empirical Evidence from Food-Delivery Platforms

Food waste is a severe economic and social problem. Restaurants contribute significantly to food waste because they face the classic trade-off between speed of service and leftover inventory, which is particularly crucial in the context of quick service restaurants (QSRs). To offer a high speed of service, QSRs pre-cook most of their food, but they can hold it only for a short time. To effectively manage this trade-off, QSRs have become increasingly reliant on demand forecasts. However, online food-delivery platforms that connect restaurants, riders/drivers, and consumers are growing in popularity, and it is unclear how the growth of food-delivery platforms impacts the ability of restaurants to accurately forecast their demand. We empirically investigate the impact of food-delivery platforms on the demand forecast error in QSRs and analyze the underlying mechanism. We find that as customers become increasingly dependent on food-delivery platforms, QSR demand becomes harder to forecast. We also find that the majority of the increase in overall forecast error is due to an increase in the error associated with the demand pattern and a smaller portion is due to error in forecasting demand amplitude. Based on our results, we offer suggestions for QSRs on how to manage their relationship with food-delivery platforms to decrease their forecast error and increase operational effciency.

Distributed energy-reserve Co-Optimization of electricity and natural gas systems with multi-type reserve resources

Abstract An energy-reserve co-optimization model for electricity and natural gas systems with multi-type reserve resources is proposed. Considering the reserve capacity provided by generating units is difficult to meet the large-scale regulation requirements caused by wind power, load demand forecast errors and unexpected contingencies, a reserve scheme with multi-type reserve resources including generating units, energy storage devices, and interruptible loads is designed. Besides, the transmission capability of reserve capacity in the electricity and gas networks when facing forecast errors and contingencies is considered in the proposed model. To protect information privacy of different energy systems, an improved alternating direction method of multipliers (ADMM) is adopted to solve the energy-reserve co-optimization model. The feasibility and effectiveness of the proposed model and method are verified on a modified IEEE 24-bus power system and 6-node natural gas systems.

정상상황에서의 확률분포를 고려한 운영예비력 소요량 추정

In this study, we tried to estimate the future operational reserves required for steady state in year 2030 using the probabilistic approach. To get the probability distribution of the reserve capacity requirement, we used the nonparametric estimation method, kernel density estimation. As the result, the frequency control reserve capacity is required to be about 800MW based on 2030, and the reserve level that needs to be secured due to the difference between 1 hour dispatch plan and 5 minutes actual load in the intraday market in 2030 is estimated about 1,400MW. In addition, the reserves that needs to be considered when unit commitment planning on the day ahead should include the PV generation forecast error distribution and the distribution of demand forecast errors also. We were estimated them by using the convolution, and we could get the result the operational reserves about 5,000∼ 7,500MW when establishing the day ahead unit commitment plan by 2030.

Application of a new fast analytical approach for post-contingency state estimation of “N-1” transmission line outage in online vulnerability assessment of large-scale transmission networks with wind power resources

In this paper, a new analytical method is presented for calculation of the “N-1″ transmission line outage post-contingency states of the power system, without any needs to iterate the calculation process, which makes the method fast enough for online vulnerability assessment under N-1 Transmission line outages for large-scale transmission networks with wind power resources. The proposed method is based on the proposed pretending line outage principle and a pre-contingency Taylor Series Approximation. Moreover, an adaptive technique is proposed for selection of the number of critical lines. First, various test systems including IEEE 9-, 39-, 57-, 118-, 300-, and 2746-bus are used for investigation of the performance of the proposed method in terms of accuracy and computational time. Moreover, the ability of the proposed method in solving the Online Vulnerability Assessment under the uncertainty of the operating condition, caused by wind power resources and load demand forecast error, is investigated in IEEE 118-bus and IEEE 300-bus test systems. Finally, the simulation results are presented to verify that the proposed approach has sufficient speed and accuracy to be used in online operational vulnerability assessments.

The Folly of Forecasting: The Effects of a Disaggregated Demand Forecasting System on Forecast Error, Forecast Positive Bias, and Inventory Levels

ABSTRACTPeriodic demand forecasts are the primary planning and coordination mechanism within organizations. Because most demand forecasts incorporate human judgment, they are subject to both unintentional error and intentional opportunistic bias. We examine whether a disaggregation of the forecast into various sources of demand reduces forecast error and bias. Using proprietary data from a manufacturing organization, we find that absolute demand forecast error declines following the implementation of a disaggregated forecast system. We also find a favorable effect of forecast disaggregation on finished goods inventory without a corresponding increase in costly production plan changes. We further document a decline in positive forecast bias, except for products whose production is limited owing to scarce production resources. This implies that disaggregation alone is not sufficient to overcome heightened incentives of self-interested sales managers to positively bias the forecast for the very products that an organization would like to avoid tying up in inventory.Data Availability: Data are the property of the research partner and may not be redistributed by the authors.

Impact of the COVID-19 Pandemic on the U.S. Electricity Demand and Supply: An Early View From Data.

After the onset of the recent COVID-19 pandemic, a number of studies reported on possible changes in electricity consumption trends. The overall theme of these reports was that “electricity use has decreased during the pandemic, but the power grid is still reliable”—mostly due to reduced economic activity. In this paper, we analyze electricity data until the end of May 2020, examining both electricity demand and variables that can indicate stress on the power grid. We limit this study to three states in the U.S. California, Florida and New York. The results indicate that the effect of the pandemic on electricity demand is not a simple reduction, and there are noticeable differences among regions analyzed. The variables that can indicate stress on the grid (e.g., daily peak and trough of the hourly demand, demand ramp rate, demand forecast error, and net electricity interchange) also conveyed mixed messages: some indicate an increase in stress, some indicate a decrease, and some do not indicate any clear difference. A positive message is that some of the changes that were observed around the time stay-at-home orders were issued appeared to revert back by May 2020. A key challenge in ascribing any observed change to the pandemic is correcting for weather as it can be challenging to accurately define it for large geographic regions. We provide a weather-correction method, apply it to a small city-wide area in North Central Florida, and discuss the implications of the estimated changes in demand. The results indicate that a 10% (95% CI [2%, 18%]) increase in electricity demand is likely to have occurred due to COVID-19 for the city analyzed.

Perishable Inventory Management Using GA-ANN and ICA-ANN

We have developed a multi-objective multi-product inventory management model for perishable products, focusing on the inventory management of veterinary drugs. This model minimises holding, shortage, and expired costs and also demand forecast error simultaneously. The number of expired and shortage drugs can be calculated for each period using this model. Data from three types of veterinary drugs have been collected from a distribution centre (DC). In this research, multi-layer perceptron (MLP) neural network is used to forecast the demand and genetic algorithm (GA) and imperialist competitive algorithm (ICA) are used to solve and find satisfactory solutions. In this research, artificial neural network (ANN) is combined with the two above-mentioned algorithms to solve the problem. The results show that the proposed model can find high-quality solutions because it reduces inventory costs and forecast errors in the DC. Finally, the results of combining ANN with each of the algorithms were compared and it was concluded that the combination of ANN and ICA produced better solutions.

SVR 기반 단기 전력수요 예측 알고리즘 개발

In recent years, precise power demand has become more important due to the constant increase in demand for electricity and abnormal weather conditions. Due to the abnormal weather conditions such as heat wave during summer season, electric power demand forecast error is increasing due to increase of heating and cooling load and this leads to economic losses due to the lowering of the stability of the power system and the increase in the cost of additional power purchase. In this paper, we propose a short-term power demand prediction algorithm based on Support Vector Regression which is an artificial intelligence technique. We implemented a short-term prediction model based on public data through correlation analysis such as temperature, time, and past electricity demand. In order to analyze the model performance, we analyzed the kernel function and slack variable value which are SVR parameter values. Finally, cross validation was performed to solve the overfitting problem.

Stochastic optimal energy management system for RTG cranes network using genetic algorithm and ensemble forecasts

Abstract In low voltage networks, Energy Storage Systems (ESSs) play a significant role in increasing energy cost savings, peak reduction and energy efficiency whilst reinforcing the electrical network infrastructure. This paper presents a stochastic optimal management system based on a Genetic Algorithm (GA) for the control of an ESS equipped with a network of electrified Rubber Tyre Gantry (RTG) cranes. The stochastic management system aims to improve the reliability and economic performance, for given ESS parameters, of a network of cranes by taking into account the uncertainty in the RTGs electrical demand. A specific case study is presented using real operational data of the RTGs netwrok in the Port of Felixstowe, UK, and the results of the stochastic control system is compared to a standard set-point controller. In this paper, two forecast data sets with different levels of accuracy are used to investigate the impact of the crane demand forecast error in the proposed ESS control system. The results of the proposed control strategies indicate that the stochastic management system successfully increases the electric energy cost savings, the peak demand reductions and successfully outperforms a comparable set-point controller.

Can demand forecast accuracy be linked to airline revenue?

Since accurate demand forecasts are a key input to any airline revenue management system, it is reasonable to assume that an improvement in demand forecast accuracy would lead to increased revenues. However, this relationship has often been called into question. Past work has not conclusively proven that more accurate demand forecasts lead to higher revenue, causing researchers and practitioners to debate whether the concept of demand forecast accuracy itself is “myth or reality.” In this paper, we demonstrate that it is possible to consistently link demand forecast accuracy to airline revenue. After discussing why traditional demand forecast error metrics have struggled to demonstrate this relationship, we evaluate a novel conditional demand forecast error metric which compares demand forecasts to historical bookings conditional on the set of fare classes that were open at the time of booking. We prove under some mild assumptions that minimizing conditional demand forecast error will maximize revenue under any fare structure and customer choice behavior. These theoretical findings are supported by simulations in both a simple, single-leg model and in a complex multiple-airline network in the Passenger Origin–Destination Simulator. We find that price elasticity parameter bias of ± 10% can reduce revenues by up to about 1%, while price elasticity parameter bias of ± 20% can reduce revenues by up to 4%. We close by discussing the implications of the findings for revenue management practitioners.

The value of flexibility in power markets

The concept of flexibility is not one you find in standard microeconomics textbooks, yet it already plays a major role in the remuneration of the resources that generate and consume electricity every day and is likely to play an even larger role with the penetration of large intermittent renewable capacities. In this paper we attempt to quantify the net revenues that can be captured by a flexible resource able to react to the short term price variations on the day-ahead and intraday markets in Germany. We find that the difference between day-ahead and intraday revenues for a flexible resource has been increasing (although the profitability has been decreasing on both markets). This difference is more pronounced once 15 mn price variations can be captured by a flexible resource. The net revenues from the local 15 mn auction (which is held 3 h after the hourly “coupled” day-ahead auction) are more than eight times higher than the day-ahead hourly auction but below the net revenues that can be captured with the high prices from the continuous market. The results of the backward-looking empirical estimations allow us to distinguish and quantify two components of flexibility: (1) the “immediacy” value as we are approaching real-time and the urgency of the delivery increases (this value is revealed during the continuous intraday process and is highly linked to the stochastic nature of power supply and demand (i.e., wind/solar forecasts, forced outages of thermal generation,…) forecast error risk), and (2) the “ramping capability” component based on the technical characteristics as a resource can react to variations of shorter granularity (15 mn vs. 60 mn). We model and quantify the ramping capability component using a geometric brownian motion with jumps.

Day-ahead thermal and renewable power generation scheduling considering uncertainty

Abstract This paper proposes a multi-objective optimization (MOO) based optimal day-ahead scheduling of thermal and renewable (wind and solar photovoltaic) power generation problem considering the intermittent/uncertain nature of renewable energy sources, load demands and thermal generators. System operating cost (i.e., cost of thermal, wind, solar PV and battery powers), reliability and emission cost are considered to be optimized simultaneously. The uncertainties due to generator outages, wind, solar PV and load demand forecast errors are incorporated in the proposed optimization problem using Expected Unserved Energy (EUE) and Loss Of Load Probability (LOLP) reliability indices. In the proposed approach, the amount of spinning reserves (SRs) required are scheduled based on the desired level of system reliability. The proposed multi-objective optimization problem is solved using NSGA-II algorithm. Different case studies are performed considering two or three different objective functions that may be selected by the system operator (SO) based on his/her preference. The simulation results obtained on a sample test system validate the benefits of solving the hybrid power system scheduling problem as a transparent and realistic MOO problem considering the uncertainty.

Energy Market Design for Renewable Resources: Imbalance Settlements and Efficiency–Robustness Tradeoffs

This paper examines optimal energy market bidding behavior for stochastic resources when energy imbalance payments are determined endogenously from market clearing conditions. Some electricity markets require wind producers to settle day-ahead offers based on their real-time energy production. This is usually justified by economic and cost-causation principles, without considering the effects of these policies on curtailment and system balancing costs. Under these real-time settlement requirements, demand and wind forecast errors impact the effective penalty price, so strategic behavior affects the penalty price paid by producers. This paper formulates a two-stage game where producers place energy bids in the day-ahead market and then choose a curtailment level in real-time, and it derives the symmetric subgame perfect equilibrium strategy for producers. It details how supply curve convexity and resource correlation can lead wind producers to bid conservatively even when they are not risk averse. The results show that market-based pricing mechanisms improve tradeoffs between system efficiency and mean squared bid error as compared to the case where penalty levels are determined by alternative policies, suggesting additional benefits of market-based penalty prices beyond those previously studied.

Robust gateway placement in wireless mesh networks

Wireless mesh networks (WMNs) are communication networks that provide wireless Internet access over areas with limited infrastructure. Each node in a WMN serves several clients in its coverage area and transfers their traffic over wireless media to a few gateway nodes that have wired connections to the Internet. In this paper, we consider the Internet gateway placement (IGP) problem along with operational problems such as routing and wireless transmission capacity allocation. To eliminate wireless interference, we adopt the spatial reuse time division multiple access (TDMA) method, in which wireless transmissions are scheduled to occur at different time slots. We also employ destination-based single path routing for its ease of implementation. We present two mixed integer linear programming (MILP) formulations, both of which jointly determine the minimum number of gateway nodes needed to support forecasted demand, the locations of these gateway nodes, the routing trees, and the time slot allocations to wireless links. These formulations differ in the flow constraints. We also present a set of valid inequalities for the formulation with the multi-commodity flow constraints. In most cases, the solution to the IGP problem is not unique. Therefore, we also introduce a local search algorithm to select the most robust solution against any demand forecast errors. On example networks, we compare the proposed formulations with and without the valid inequalities in terms of the exact solution performances and the linear programming (LP) relaxations. We also demonstrate our local search algorithm to improve robustness against forecast errors on these example networks.

Inventory misplacement and demand forecast error in the supply chain: profitable RFID strategies under wholesale and buy-back contracts

We investigate RFID adoption strategies under wholesale price and buy-back contracts in a supply chain with one manufacturer and one retailer who faces inventory misplacement and demand forecast error. RFID can alleviate the misplacement problem, and can reduce demand forecast error by shortening order lead time. By a newsvendor model, we characterise the optimal contract terms in the supply chain without and with RFID adoption, respectively. We further analyse how the contract terms depend on RFID-related parameters (e.g. salable rate and demand forecast error). We find that both without and with RFID, the wholesale price contract will lead to the double marginalisation problem, while the buy-back contract can coordinate the supply chain. We show that the supply chain adopts RFID if and only if the tagging cost is below a threshold; the threshold is in negative correlation to the demand forecast error. The supply chain is more willing to adopt RFID under the buy-back contract than under the wholesale price contract. RFID adoption can sometimes lessen the double marginalisation problem under the wholesale price contract, improving the supply chain efficiency. A smaller RFID tagging cost or a reduced forecast error do not necessarily lead to higher supply chain efficiency.

A multi-timescale hybrid stochastic/deterministic generation scheduling framework with flexiramp and cycliramp costs

Flexible ramping products (flexiramp), provided by entitled resources to meet net demand forecast error, are the underpinning for the accommodation of the substantial uncertainties associated with variable wind power. This paper proposes an enhanced flexiramp modeling approach, cast in a hybrid stochastic/deterministic multi-timescale framework. The framework employs a chance-constrained day-ahead scheduling method, as well as deterministic scheduling on intra-hourly basis (real-time scheduling), to allow optimal procurement planning of the flexiramp products in both timescales. A stepwise and piecewise demand price curve is also proposed to calculate the flexiramp surplus procurement price. Non-generation resource (NGR), referring to energy storage, is implemented to provide extra flexibility. Additionally, cycling ramping cost (cycliramp), introduced to model operational and maintenance costs and reduce the wear and tear of generators, is also included as a penalty. Numerical tests are conducted on 6-bus and 118-bus systems. Results demonstrate the merits of the proposed scheduling model as well as the effects of flexiramp and cycliramp costs in the multi-timescale scheduling.