Unsold Items Research Articles

Risk management for barter exchange policy under retail industry

Nowadays, barter exchange has become growingly popular in the national and global industries as an alternative to excessive inventory transfers. Many companies exchange their surplus products on barter platforms for products they need without using money. In a traditional supply chain, the retailer holds unsold products that hinder the supply chain's profitability from reaching its maximum level. This paper solves this issue by proposing a stochastic model of two players (single manufacturer, single retailer) with trade-credit (delayed payment) and barter exchange policy under a supply chain management. In this work, to entice the retailer and increase sales, the manufacturer grants a credit payment facility to the retailer on the items ordered for a specified period, and the manufacturer does not charge any interest on the outstanding amount during this credit period. The concept of a credit period raises the possibility of default risk. In this case, some interest is charged. Several investments are made here to diminish setup and ordering costs and improve the product quality of the system. This work focuses on the flexible production to manage the demand uncertainty and the marginal reduction technology to lessen carbon emissions that occur during the production and inventory holding. In contrast, a retailer can exchange unsold items in the barter market for its required products, which is extensively discussed in this study. Finally, the maximum profit is assessed in terms of credit period, investments, quality improvement, and production rate. The result numerically and graphically proves a huge impact of the barter platform for any business industry on overstock transfers, conserving cash, managing unpredictable demand, and reaching the maximum profit. Moreover, the significant finding is observed in the proposed work that the idea of the flexible production, barter exchange policy, and several investments increase the system profit up to 50.55%.

Store Analysis Using Latent Representation of Robust Variational Autoencoder Based on Sales History Data

This study proposes an analytical model based on the robust variational autoencoder (RVAE) to analyze store characteristics and further formulate item strategies at the store level for future reduction of the loss of unsold items. The data of the display item vector of each store characterizes the sales strategy of the store. However, it is sparse because the number of items displayed in each store accounts for only approximately 10% of the total number of item brands sold in all the stores. To manage this sparsity, a simple autoencoder model is inappropriate because of the large variation in input data, which is as a result of the store characteristics of selling unique products. Therefore, this study employs the RVAE to analyze the store characteristics of the listed items. Here, the latent representation of an RVAE is usually an output from an estimated probability distribution in the middle layer of the trained network. Generally, the similarity of the latent representations of two input data is evaluated using the sample data of the probability distributions. This study proposes a model to calculate the distance between probability distributions without sampling, for valid estimations of the similarities between latent representations. It proposes a method to use the reconstruction error obtained through the RVAE that enables the detection of stores whose trends differ significantly from those of other stores. Additionally, the proposed method can detect groups of stores with similar trends. In this study, the proposed model was applied to an actual dataset, after which its effectiveness was verified.

APPLICATION OF THE FITRADEOFF METHOD IN A PORTFOLIO PROBLEM IN THE CONTEXT OF REVERSE LOGISTICS FOR WHOLESALE

The study investigates options to support the reverse logistics for wholesalers in a company. This big company, which do not have the name presented in this paper aims to improve the process of returning unsold items. At the moment, the process contained several rework-related losses. Hence, in this study two methodologies have been applied. The VFT has been used to structure the problem considering the multiple objectives presented by the Decision-Maker, in special those that are considered by sporting multinational. Also, the FITradeoff method has been used to identify the alternatives that should compose the portfolio of options concerning the reverse logistics. The portfolio problematic has been applied since it has been desired to obtain the subset of option to support the problem, constrained by monetary resources. Using the VFT, ten alternatives which were evaluated in ten criteria have been defined, and support by the FITradeoff method. As result, the alternatives five alternatives have been selected to compose the portfolio.

Explainable product backorder prediction exploiting CNN: Introducing explainable models in businesses

Due to expected positive impacts on business, the application of artificial intelligence has been widely increased. The decision-making procedures of those models are often complex and not easily understandable to the company’s stakeholders, i.e. the people having to follow up on recommendations or try to understand automated decisions of a system. This opaqueness and black-box nature might hinder adoption, as users struggle to make sense and trust the predictions of AI models. Recent research on eXplainable Artificial Intelligence (XAI) focused mainly on explaining the models to AI experts with the purpose of debugging and improving the performance of the models. In this article, we explore how such systems could be made explainable to the stakeholders. For doing so, we propose a new convolutional neural network (CNN)-based explainable predictive model for product backorder prediction in inventory management. Backorders are orders that customers place for products that are currently not in stock. The company now takes the risk to produce or acquire the backordered products while in the meantime, customers can cancel their orders if that takes too long, leaving the company with unsold items in their inventory. Hence, for their strategic inventory management, companies need to make decisions based on assumptions. Our argument is that these tasks can be improved by offering explanations for AI recommendations. Hence, our research investigates how such explanations could be provided, employing Shapley additive explanations to explain the overall models’ priority in decision-making. Besides that, we introduce locally interpretable surrogate models that can explain any individual prediction of a model. The experimental results demonstrate effectiveness in predicting backorders in terms of standard evaluation metrics and outperform known related works with AUC 0.9489. Our approach demonstrates how current limitations of predictive technologies can be addressed in the business domain.

LINEAR REGRESSION FOR DISCOUNTING PRESENTATION RECOMMENDATIONS (Kaggle Dataset)

In the business of selling goods, there must be goods that do not sell well and sell well. How to make unsold items sell well by giving customers a discount or discount strategy. The goal is to provide discounted prices to attract customers' attention and increase sales turnover. Prediction to give the right discount presentation is needed in the discount strategy. How to determine discount prediction using linear regression method, looking for line equations by training data taken from Kaggle.com. The data were trained to find the constants and coefficients of the independent variables. The resulting line equation will be tested to determine the discount prediction that will be given and calculate the error value. The data used is painting sales data taken from Kaggle.com, a total dataset of 1525, and a prediction calculation is made with 225 lines of test data. The input variable used is the price of the painting, and the output variable is the discount presentation. The process of evaluating the prediction results by calculating the MAE value, the difference between the actual data and the predicted data, the MAE value from the trial is 0.108.

Design of Machine Learning Algorithm for Tourism Demand Prediction.

Unused hotel rooms, unused event tickets, and unsold items are all examples of wasted expenses and earnings. Governments require accurate tourism demand forecasting in order to make informed decisions on topics such as infrastructure development and lodging site planning; therefore, accurate tourism demand forecasting becomes vital. Artificial intelligence (AI) models such as neural networks and security violation report (SVR) have been used effectively in tourist demand forecasting as a result of the fast advancement of AI. This paper constructs a tourism demand forecasting model based on machine learning on the basis of the existing forecasting model research. The completed work is as follows: (1) It introduces a large number of domestic and foreign literatures on tourism volume forecasting and proposes the research content of this paper. (2) It is proposed to stack the long short-term memory- (LSTM-) based autoencoders deeply, by adopting a hierarchical greedy pretraining method to replace the random weight initialization method used in the deep network and combining this pretraining stage and fine-tuning network together to form the SAE-LSTM prediction model for improving the performance of deep learning models. (3) This paper uses the monthly search engine strength data of city A's monthly tourist volume and its related influencing factors as the data set; processes the data set to make the model adapt to the data input; uses mean absolute error (MAE), root mean square error (RMSE), MAPE, and other model evaluation indicators; and uses LSTM and the constructed SAE-LSTM model to conduct comparative experiments to predict the number of tourist arrivals in four years. The prediction results of the models proposed in this paper are better than those of the LSTM model. According to the experimental results, the superiority of the proposed LSTM-based unsupervised pretraining method is demonstrated.

The impact of buyback support on financing strategies for a capital-constrained supplier

In this paper, we investigate the optimal operational decisions and financing strategies of a supply chain with one financial-constrained supplier and one reputable retailer. In particular, two pre-shipment financing mechanisms, buyback support buyer-backed purchase order financing (BSBPOF) and buyback support advance payment discount (BSAPD), are chosen to ease the supplier's financial pressures. Both financing methods indicate that the supplier promises to buy back the unsold items after the sales season, but the difference is that the former means that the financial institution lends money to the supplier which the retailer needs to guarantee, while the latter means that the retailer pays in advance. Our research shows that: 1) the difference between the buyback price of the unsold product and its salvage value has an impact on the optimal operational decisions (such as the supplier's capacity level, the discount rate under BSAPD, and the financial institution's interest rate under BSBPOF); 2) when both financing instruments are available, if the buyback price is above a certain level and the difference between the buyback price and the salvage value is less than a certain range, the financing equilibrium is BSBPOF. Otherwise, BSAPD is preferred. Furthermore, the results provide managerial insights on how to address the operational and financing issues, and relevant conclusions are drawn by extending the model to consider stockout costs.

A distribution-free newsvendor model considering environmental impact and shortages with price-dependent stochastic demand.

In today's competitive and volatile market, demand prediction for seasonal items is a challenging task. The variation in demand is so quick that the retailer cannot face the risk of understocking or overstocking. Unsold items need to discarded, which has environmental implications. It is often difficult to calculate the effects of lost sales on a firm's monetary values, and environmental impact is not a concern to most businesses. These issues concerned with the environmental impact and the shortages are considered in this paper. A single-period inventory mathematical model is formulated to maximize expected profit in a stochastic scenario while calculating the optimal price and order quantity. The demand considered in this model is price-dependent, with several emergency backordering options to overcome the shortages. The demand probability distribution is unknown to the newsvendor problem. The only available demand data are the mean and standard deviation. In this model, the distribution-free method is applied. A numerical example is provided to demonstrate the model's applicability. To prove that this model is robust, sensitivity analysis is performed.

Machine Learning Outperforms Classical Forecasting on Horticultural Sales Predictions

Forecasting future demand is of high importance for many companies as it affects operational decisions. This is especially relevant for products with a short shelf life due to the potential disposal of unsold items. Horticultural products are highly influenced by this, however with limited attention in forecasting research so far. Beyond that, many forecasting competitions show a competitive performance of classical forecasting methods. For the first time, we empirically compared the performance of nine state-of-the-art machine learning and three classical forecasting algorithms for horticultural sales predictions. We show that machine learning methods were superior in all our experiments, with the gradient boosted ensemble learner XGBoost being the top performer in 14 out of 15 comparisons. This advantage over classical forecasting approaches increased for datasets with multiple seasons. Further, we show that including additional external factors, such as weather and holiday information, as well as meta-features led to a boost in predictive performance. In addition, we investigated whether the algorithms can capture the sudden increase in demand of horticultural products during the SARS-CoV-2 pandemic in 2020. For this special case, XGBoost was also superior. All code and data is publicly available on GitHub: https://github.com/grimmlab/HorticulturalSalesPredictions.

A study on green supply chain under capital constraint considering time-varying salvage value

To reduce environmental pollution and promote sustainable development, more and more suppliers are committed to producing more environmentally friendly products such as green electrical appliances through a green supply chain (GSC) system. However, some suppliers are often limited by a lack of funds in the production and supply process. In this paper, buyer-supported purchase order financing (BPOF) and advance payment discount (APD) are adopted to help the supplier successfully produce green products in a GSC system consisting of a financially constrained supplier and a reputable retailer. Moreover, the salvage values of unsold inventory in most traditional models are fixed, but in real life, the salvage value tends to be time dependent. Therefore, how the time-varying salvage value affects the operation and financing decisions as well as the profit risks is studied. We find that the clearance time of the unsold items affects the optimal order quantity and the supplier's discount rate and the financial institution's optimal interest rate. In addition, the financing equilibrium is BPOF under certain conditions, and the profit risks of the retailer and the entire supply chain are increasing with the order quantity and the clearance time. Finally, our results are verified through numerical analysis.

Transfer of Risk in Supply Chain Management with Joint Pricing and Inventory Decision Considering Shortages

This study is the first to consider a distribution-free approach in a newsvendor model with a transfer of risk and back-ordering. Previously, in many articles, discrete demand is considered. In this model, we consider a newsvendor selling a single seasonal item with price-dependent stochastic demand. Competition in markets has forced the retailer and manufacturer to coordinate in decentralized supply chain management. A coordination contract is made between a retailer and manufacturer to overcome the randomness of demand for a short-life-cycle product. The retailer pays an additional amount per product to transfer the risk of unsold items. The manufacturer bears the cost of unsold products from the retailer. Shortages are allowed with back-ordering costs during the season. The distribution-free model is developed and solved with only available demand data of mean and standard deviation. Stackelberg’s game approach is used to calculate the optimal ordering quality and price. This model aims to maximize expected profit by optimizing unit selling price and ordered quantity through coordination. To illustrate that the model is robust, numerical experiment and sensitivity analyses are conducted for both decentralized and centralized supply chain management. For applicability of the model in the real-world business scenario, managerial insights are provided with sensitivity analysis.

Assessment of Carbon Footprint for the Textile Sector in France

Global warming represents a major subject on all society levels including governments, economic actors and citizens. The textile industry is often considered a polluting activity. In this project, French textile manufacturers sought to quantify the carbon footprint (CF) of sold clothes and household linen using Life Cycle Assessment in France for the purpose of reducing it to meet the constraints of Paris Agreement by 2050. First, manufacturers calculated the carbon footprint of 17 clothes and household linen products and established alternative scenarios for four production routes. Secondly, they modeled the supply of the upstream sector through different countries. Based on imports of textile products, their calculated CF for one French person reaches 442 kg of CO2eq/year. Means of action to reduce this carbon footprint by a factor of 6 (74 kg of CO2eq/person/year for textiles) are calculated and are the following: installing the most energy-intensive production processes in a country with a low carbon electricity mix, avoiding unsold goods, implementing eco-design approaches and enhancing the value of end-of-life products with reuse or recycling. Therefore, CF for textiles per capita is reduced to 43 kg CO2eq/year which goes beyond the objectives of Paris Agreement and facilitates carbon neutrality in the textile sector. The first priority for reducing the French carbon footprint of clothes and household linen would be to locate textile production in countries with (i) low carbon electricity, (ii) to reduce unsold items, and (iii) to elaborate ecodesign of product including circular economy.

Optimal ordering and disposal decisions for products with a fixed shelf life

In this paper, motivated by the current increasing interest and action on food waste reduction, inventory decisions of a retailer who deals with a product that has a fixed shelf life are studied. Being a common strategy of many retail stores, we assume that at a specific time instant, close to the expiration date, a price markdown is offered in order to increase demand. However, at the same time, due to customers’ attention to the freshness of the product, the demand becomes a decreasing function with respect to the time remaining before the expiration date. In accordance with the European Union food donation guidelines, we assume that if at the end of the reorder interval unsold items remain that have not exceeded their expiration date, they can be donated to non-profit organizations for human consumption. The donated products can generate direct revenue from tax deductions and indirect revenue by increasing the company’s reputation and gain of goodwill from the customers. If the unsold items have expired, they can be sold at a salvage price to the livestock market. The aim of our model is to determine the reorder interval, the time instant to markdown the product’s initial selling price and the quantity that will be donated or sold to the livestock market so that the profit of the system is maximized. Closed form solutions are obtained, which depend on specific parametric conditions, providing managerial insights.

Single cycle supply chain coordination model for fuzzy stochastic demand of perishable items

Market demand of the perishable bakery products are volatile. Hence, demand of such items are random variable. The most appealing demand distribution of such items is lognormal demand distribution. Using historical data, the parameters of lognormal distribution are obtained. But, there is a problem of overestimation or underestimation of the parameters. Fuzzy triangular numbers are used to overcome such problems. Single period (newsboy) inventory model is used to obtain an optimal order quantity. The retailers and manufacturers profit is obtained by using the supply chain coordination model. At the end of day, all the remaining items rotted. Hence, there is no gain from unsold items. Also, the some items are defective at the time of purchase. Thus, loss sharing strategy is proposed between manufacturers and retailers. All of these aspects are incorporated in the supply chain coordination model. The data is collected from Koregaon Park, Pune market, India to illustrate the methodology.

Invest in Information or Wing It? A Model of Dynamic Pricing with Seller Learning

Pricing products such as used cars, houses, and artwork is often challenging, because each item is unique, and the seller, ex ante, lacks information about the demand for individual items. This paper develops a dynamic pricing model for products with significant item-specific demand uncertainty, in which a forward-looking seller learns about the item-specific demand through an initial assessment, as well as during the selling process. The model demonstrates how seller learning, through several mechanisms, can lead to the commonly observed downward trend in the prices of individual items. These mechanisms include the seller’s optimal adjustment of prices over time to account for the dynamic adverse selection of unsold items and the diminishing option value in future learning. The model is estimated using novel panel data of a leading used-car dealership. Counterfactual experiments show that the value of learning in the selling process is $203 per car. Conditional on subsequent learning in the selling process, the initial assessment further improves profit per car by $139. With the dealer’s net profit per car being about $1,150, these estimates suggest a potentially high return to taking an information-based approach toward pricing products with item-specific demand uncertainty. This paper was accepted by Juanjuan Zhang, marketing.

Single stage fuzzy supply chain model with Weibull distributed demand for milk commodities

The newsvendor model is very simple, but it is very important to study problems of supply chain coordination. This model is also called a single-cycle inventory model. In a short cycle, it is very difficult to forecast the demand. The effect of good prediction reduces inventory shortage, increases the profit and enhances competitiveness. This type of product is characterized by strong uncertainty, that means demand always changes. This creates challenges for decision makers. For newly developed products, it is very difficult to obtain the statistical demand distribution of a product. Here, we consider the demand of the milk products, which are fluctuating in nature. Hence, demand is a random variable. It is assumed that demand follows Weibull distribution. Also, there is some uncertainty in demand. It leads to the over estimation or under estimation of the parameters of the distribution. Fuzzy triangular numbers are used to obtain the uncertain demand. The lifetime of the items are very short. Thus, unsold items are sent back to the manufacturers. The retailers and manufacturers profit is obtained under decentralized and centralized supply chain. Also,the proposed single period (newsboy) inventory model is used to obtain an optimal order quantity. The proposed methodology is illustrated for milk product from Aundh market Pune, India.

Seasonal Inventory Decisions (Single-Period Inventory Models)

<p>One of the dilemmas facing many retailers is how to handle seasonal goods, such as sarung dress or batik dress, biscuits and syrups during Idul Fitri celebration. Often they cannot be sold at full markup next year because of changes in styles, and expiry dates for foods and beverages. Furthermore, the lead time can be longer than the selling season, allowing no second chance to rush through another order to cover unexpectedly high demand. At the end of the period the product has either sold out or there is a surplus of unsold items to sell for a salvage value. The single-period inventory models are used in situations involving seasonal or perishable items that cannot be carried in inventory and sold in future periods.<br />This type of situation is often called the newsboy problem. If the newspaper seller does not buy enough newspapers to resell on the street corner, sales opportunities are lost. If the seller buys too many newspapers, the overage cannot be sold because nobody wants yesterday’s newspaper.<br />At this opportunity, we will use the single-period inventory models decisions to create an optimal order quantity decision with the highest expected payoff. We will be able to use this decision process for all such items over many selling seasons to maximize profits.<br />However, it is not a foolproof, and it can result in an occasional bad outcome. It will depend on the situation, if the probability of demand follows the demand probability planning, then you will be able to get the best decision to prepare and execute an order.<br />Kata kunci:<br />Seasonal Inventory, Single-Period Inventory Models, Probability of Demand.</p>

Pareto Improving Supply Chain Coordination Under a Money-Back Guarantee Service Program

Money-back guarantee (MBG) is a well-known service offered by many retailers. Under the MBG policy, a purchased item by a consumer can be returned for a full refund. In this paper, we explore a supply chain (SC) system comprising of one retailer, which offers the money-back guarantee policy and faces a stochastic demand. With a given supply contract offered by the supplier, the retailer makes decisions on the order quantity and the market coverage of the money-back guarantee service. With reference to real-world practices, we examine three models, namely, (1) the traditional wholesale pricing model without the buyback contract, (2) the model with a buyback contract (between the supplier and the retailer) for unsold items, and (3) a dual-buyback (DB) contract model for both unsold and consumer returned items. We find that using the buyback contract alone for unsold items cannot achieve Pareto improving supply chain coordination, whereas the DB contract can. Numerical findings also indicate that the DB contract can lead to an expansion of the money-back guarantee service program to cover more customers. This paper’s findings support the implementation of the DB contract, which is commonly seen in the industry, for efficient supply chain management in the presence of the money-back guarantee service program.

Take-back agreements in the perspective of food waste generation at the supplier-retailer interface

Food waste must be minimised to make food supply chains sustainable. This is especially relevant since food waste valorisation measures, such as energy recovery, have limited possibilities to fully recover the resources invested in food production. However, waste minimisation is costly when it includes new infrastructure or technology. Policy measures, on the other hand, can provide a low-cost option. Food rejection practices in supermarkets, such as take-back agreements (TBA), have long been identified as risk factors for food waste generation at the supplier-retailer interface, but given the relational, and often discreet, nature of these agreements, there is little evidence of their impact. In this study we provide, concrete evidence of different rejection practices. This is done by studying three types of food chains – those for bread, fresh fruit and vegetables, and milk – with different rejection practices in Sweden. Based on a combination of primary company information and stakeholder interviews, we found that a full TBA is in operation for bread. The retailer only pays for bread that is sold and any bread left unsold three days before the best-before date is returned to the supplier. For fresh fruit and vegetables, only goods of ‘inadequate’ quality are returned, but supermarkets have sole rights of determination on quality, posing a risk of categorising unsold fruit and vegetables as inadequate quality and returning them to suppliers. In the case of milk, suppliers take back unsold items, but only for waste management. The trend found in this study was that bread had the highest waste, and the most extensive take-back policy. Fresh fruit and vegetables had medium levels of waste, partly due to unverified rejections, while milk had a very low level of waste combined with an even lower level of rejections. It can be concluded that a food supply chain system where the direct costs of waste management or incentives for waste reduction are separated from the organisation responsible for generating the waste poses a significant risk factor in food waste generation and is therefore a potential hotspot for waste-reducing measures.

Cluster-based hierarchical demand forecasting for perishable goods

Demand forecasting is of particular importance for retailers in the context of supply chains of perishable goods and fresh food. Such goods are daily produced and delivered as they need to be provided as fresh as possible and quickly deteriorate. Demand underestimation and overestimation negatively affect the revenues of the retailer. Stock-outs have an undesired impact on consumers while unsold items need to be discarded at the end of the day. We propose a DSS that supports day-to-operations by providing hierarchical forecasts at different organizational levels based on most recent point-of-sales data. It identifies article clusters that are used to extend the hierarchy based on intra-day sales pattern. We apply multivariate ARIMA models to forecast the daily demand to support operational decisions. We evaluate the approach with point-of-sales data of an industrialized bakery chain and show that it is possible to increase the availability while limiting the loss at the same time. The cluster analysis reveals that substitutable items have similar intra-day sales pattern which makes it reasonable to forecast the demand at an aggregated level. The accuracy of top-down forecasts is comparable to direct forecasts which allows reducing the computational costs.