WO2016122292A1 - Method for forecasting the demand of non-seasonal and non-perishable products - Google Patents
Method for forecasting the demand of non-seasonal and non-perishable products Download PDFInfo
- Publication number
- WO2016122292A1 WO2016122292A1 PCT/MX2015/000015 MX2015000015W WO2016122292A1 WO 2016122292 A1 WO2016122292 A1 WO 2016122292A1 MX 2015000015 W MX2015000015 W MX 2015000015W WO 2016122292 A1 WO2016122292 A1 WO 2016122292A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- demand
- products
- data
- prediction
- forecasting
- Prior art date
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
Definitions
- the present invention corresponds to the field of computer science because it proposes a computational method of demand forecasting in supermarkets.
- the US patent application Ser. No. 6,138,103 of Cheng et al. describes a decision-making method for predicting uncertain demand.
- the Cheng system uses a matrix to represent the possible demand scenarios and their relative probabilities of occurrence, and these matrices are used to calculate a productive planning calendar based on the most likely outcome of the uncertain demand.
- Cheng like Milne, does not teach manners for the development of alternative algorithms for forecasting future demand, as well as prediction algorithms such as adjustment demand.
- U.S. Patent No. 5,459,656 to Fields et al. discloses a system that measures and stores business demand data plus a plurality of intervals of time and projects business demand for products at near future time intervals using demand curves based on percentages.
- the field system that allows the creation of a plurality of demand curves to determine demand in the near future through the use of defined functions and variables. Projections of business demand for current and near future time intervals can be reviewed in response to variations in the actual data of business demand, which you already received.
- the fields fail to produce forecasts that proactively take into account many factors that affect future demand including the variability of individual markets and the impact of product promotions.
- US Patent No. 6,032,125 to Ando describes a demand forecasting system that allows users to forecast demand based on algorithms derived from data from different time periods, including the latest months, similar periods of previous years and their combinations.
- the Ando patent does not take into account that data on past demand may come from various sources within a single organization or supply chain (for example, sales data, returns, wholesale data, etc. .).
- Ando does not teach systems or methods to derive several models of demand forecasting and the revision of those models over time.
- Figure 1 shows a graph of the standard deviation of a product.
- Figure 2 represents a flow chart of the forecasting process.
- the following method is proposed for the prediction of the demand for products in commercial establishment.
- the sales presented represent the total demand, that is, the establishment had no storage restrictions and always had unit availability, so that the sales generated represent the actual demand of the product.
- the products are not seasonal. That is, its sale does not vary substantially at a certain time, on certain days of the week or in a certain climate.
- the products are not perishable, so that their freshness does not impact their consumption.
- the first step is to calculate the correlation between products. If there is one or more products with a correlation above a limit established at the discretion (for example, .9), a correlation table is established between both products.
- a limit established at the discretion for example, .9
- the table will keep a record of the products whose behavior is observed with a correlation above a specific criterion (in this case,> 0.9). In this case, product A and C have a positive correlation of 0.96, while with product B it has a correlation of 0.90 2. Calculate the descriptive statistics of the sales and analyze the data of the first and last standard deviation.
- the sales distribution of the product is observed during the year of history. Values that fall into the first and last standard deviation are excluded from the sample.
- the intention is to use the data to train a time series model avoiding bias.
- step 2 Through an analysis of descriptive statistics of the sales that fall into the data excluded in step 2, it is determined whether there is a marked enough trend to determine a significant variation on certain days of the week or on specific weeks during the year.
- the data hosted in this period will contain the lowest and highest sales in the entire history, so it is important to consider them for the projection.
- the graph of Figure 1 shows the first (-3) and last standard deviation (3) containing the low and high peaks of the demand for a product. If we look at the mode of the days of the week or month with a sufficiently high frequency, we can take into account the specific days to determine the demand in the future.
- the hidden markov model trains with 5 hidden states and is used to predict demand for the following week or the next day. This prediction is considered normal and is the basis for the final projection, which will take into account the extraordinary sales periods.
- the increase or distribution is taken into account and applied to the projection thrown by the model.
Abstract
Factors such as unusual weather conditions, competitive incursions and demographic variations, inter alia, greatly complicate the forecasting of product demand using traditional methods. For this reason, the invention relates to a method based on Artificial Intelligence techniques for forecasting product demand, so that, as such, the establishment buys the exact inventory that it will sell, thereby reducing the loss thereof.
Description
MÉTODO PARA EL PRONÓSTICO DE DEMANDA DE PRODUCTOS NO ESTACIONALES Y METHOD FOR THE FORECAST OF DEMAND FOR NON-SEASONAL PRODUCTS AND
NO PERECEDEROS. DON'T LOSE YOU.
CAMPO TÉCNICO DE LA INVENCIÓN TECHNICAL FIELD OF THE INVENTION
La presente invención corresponde al campo de la informática pues propone un método computacional de pronostico de demanda en supermercados. The present invention corresponds to the field of computer science because it proposes a computational method of demand forecasting in supermarkets.
ANTECEDENTES BACKGROUND
Los desajustes de la demanda y la oferta son costosos para los vendedores en un mercado competitivo, porque tales desajustes suelen dar lugar a oportunidades de ventas, pérdida de ganancias, costos de agilización excesivos, perdió de mercado, y un precario servicio al cliente. Para maximizar las ventas las empresas deben predecir con exactitud el futuro de la demanda del cliente y utilizar esta información para conducir sus operaciones de negocio de la fabricación a operaciones de distribución. Esta necesidad de predicciones precisas de la demanda es especialmente importante para los que participan en el comercio electrónico debido a la facilidad con la que los compradores pueden encontrar vendedores alternativos que pueden satisfacer su demanda. Demand and supply mismatches are costly for sellers in a competitive market, because such imbalances often lead to sales opportunities, loss of profits, excessive expediting costs, lost market, and poor customer service. To maximize sales, companies must accurately predict the future of customer demand and use this information to drive their manufacturing business operations to distribution operations. This need for accurate demand predictions is especially important for those who participate in e-commerce because of the ease with which buyers can find alternative sellers that can meet their demand.
En las industrias de fabricación y distribución, el suministro de productos en respuesta al nivel actual de la demanda de los clientes con un mínimo de exceso de existencias reduce los costes de almacenamiento y los gastos de distribución y por lo tanto conduce a una reducción del precio unitario de venta de productos. Típicamente, esto también conduce a una mejora de los márgenes de beneficio. Por ello es necesario que los vendedores pronostiquen la demanda del producto de manera precisa, de tal manera que puedan determinar un plan de ventas, plan de producción y el plan de distribución de acuerdo con un pronóstico preciso de la tendencia de la demanda de los clientes. In the manufacturing and distribution industries, the supply of products in response to the current level of customer demand with a minimum of excess stock reduces storage costs and distribution costs and therefore leads to a price reduction Unitary product sales. Typically, this also leads to an improvement in profit margins. Therefore, it is necessary for sellers to forecast the demand for the product accurately, so that they can determine a sales plan, production plan and distribution plan according to an accurate forecast of the trend in customer demand. .
Los métodos convencionales de previsión de la demanda mediante el análisis de la tendencia de resultados de ventas pasadas se realizan con el objetivo de aplicar las técnicas de análisis estadístico más precisas y modelos econométricos para proporcionar la previsión más exacta posible. En estos métodos convencionales, la predicción de series de tiempo desarrolla y utiliza varios algoritmos de predicción que intentan describir el conocimiento de la tendencia de negocios y la fluctuación de los resultados de ventas como se evidencia por la historia pasada en forma de una regla. El desarrollo de tales algoritmos de predicción (así como sistemas
¡nformatizados para la utilización de tales algoritmos) es típicamente una tarea de trabajo intensivo. Conventional methods of demand forecasting by analyzing the trend of past sales results are performed with the objective of applying the most accurate statistical analysis techniques and econometric models to provide the most accurate forecast possible. In these conventional methods, time series prediction develops and uses several prediction algorithms that attempt to describe knowledge of the business trend and the fluctuation of sales results as evidenced by past history in the form of a rule. The development of such prediction algorithms (as well as systems Information for the use of such algorithms) is typically a labor intensive task.
Desafortunadamente, sin embargo, es común que las tendencias de la demanda de productos cambie en un corto ciclo de vida. Cuando este es el caso, los datos utilizados en la presentación de un pronóstico se convierte rápidamente de vieja y la precisión de la predicción disminuye. Por lo tanto, con el fin de mantener una alta precisión en la predicción, algoritmos (y puntos de datos históricos utilizados para generar los algoritmos) se deben mantener en forma permanente, así como ser capaces de ajustar sus previsiones con relativa facilidad. Unfortunately, however, it is common for product demand trends to change in a short life cycle. When this is the case, the data used in the presentation of a forecast quickly becomes old and the accuracy of the prediction decreases. Therefore, in order to maintain high prediction accuracy, algorithms (and historical data points used to generate the algorithms) must be maintained permanently, as well as being able to adjust their forecasts with relative ease.
Por lo tanto, ya que una de las cuestiones más importantes encontradas en la planificación de la producción proviene de las incertidumbres asociadas a la demanda futura de productos, un gran volumen de literatura y de esfuerzos de la industria ha tratado de abordar esta cuestión. Actualmente, sin embargo, la producción, los materiales y la planificación del transporte en base a la demanda prevista, plantea aún un reto importante. Aunque ha habido muchos estudios en el área de la teoría de planificación de la demanda, los avances alcanzados hasta ahora están basados en suposiciones simplistas o alternativamente son computacionalmente imposible para aplicación en el mundo real. Por lo tanto, hasta ahora no ha habido desarrollado una manera todavía automatizado sustancialmente flexible con el cual los vendedores en un mercado competitivo puedan satisfacer sus necesidades de previsión de la demanda. Therefore, since one of the most important issues found in production planning stems from the uncertainties associated with future demand for products, a large volume of literature and industry efforts have tried to address this issue. Currently, however, production, materials and transportation planning based on the expected demand, still poses an important challenge. Although there have been many studies in the area of demand planning theory, the advances achieved so far are based on simplistic assumptions or alternatively are computationally impossible for real-world application. Therefore, so far it has not developed a substantially flexible yet automated way in which sellers in a competitive market can meet their demand forecasting needs.
A modo de ejemplo, la solicitud de patente de EE.UU.. No. 6.049.742 a Milne et al. da a conocer una herramienta de software informático para comparar las decisiones de planificación de suministro proyectados con perfiles de demanda esperados. El sistema informático divulgada por Milne proporciona rutinas que comparan la oferta proyectadas con la demanda real con experiencia para ayudar al usuario de la herramienta de software para configurarlo para satisfacer mejor las necesidades de su negocio. Por desgracia, al igual que las metodologías de previsión de la demanda más conocidos, Milne es inflexible en que no ayuda a los usuarios desarrollar e identificar modelos mejorados mediante la comparación de múltiples modelos alternativos para varios productos dentro de los distintos mercados. By way of example, U.S. Patent Application No. 6,049,742 to Milne et al. Unveils a computer software tool to compare projected supply planning decisions with expected demand profiles. The computer system disclosed by Milne provides routines that compare the projected supply with the actual demand with experience to help the user of the software tool to configure it to better meet the needs of your business. Unfortunately, like the best known demand forecasting methodologies, Milne is adamant that it doesn't help users develop and identify improved models by comparing multiple alternative models for various products within different markets.
De manera similar, la solicitud de patente EE.UU. Ser. No. 6.138.103 de Cheng et al. describe un método de toma de decisiones para la predicción de demanda incierta. El sistema Cheng utiliza una matriz para representar los posibles escenarios de demanda y sus probabilidades relativas de ocurrencia, y estas matrices se utilizan para calcular un calendario de planificación productiva basada en el resultado más probable de la demanda incierta. Cheng, como Milne, no enseña modales para el desarrollo de algoritmos alternativos para la previsión de la demanda futura, así como algoritmos de predicción como la demanda de ajuste. Similarly, the US patent application Ser. No. 6,138,103 of Cheng et al. describes a decision-making method for predicting uncertain demand. The Cheng system uses a matrix to represent the possible demand scenarios and their relative probabilities of occurrence, and these matrices are used to calculate a productive planning calendar based on the most likely outcome of the uncertain demand. Cheng, like Milne, does not teach manners for the development of alternative algorithms for forecasting future demand, as well as prediction algorithms such as adjustment demand.
Además, la patente de EE.UU.. No. 5.459.656 de Fields et al. da a conocer un sistema que mide y almacena los datos de demanda de negocios más una pluralidad de intervalos de
tiempo y proyecta la demanda empresarial de productos en intervalos de tiempo casi futuros utilizando las curvas de demanda basados en porcentajes. El sistema de campos que permite la creación de una pluralidad de curvas de demanda para determinar en un futuro cercano la demanda mediante el uso de funciones y variables definidas. Proyecciones de la demanda de negocios para los intervalos actuales y del futuro próximo de tiempo pueden ser revisados en respuesta a las variaciones en los datos reales de la demanda empresarial, que ya recibió. Los campos, sin embargo, no logra producir previsiones que proactivamente tengan en cuenta muchos factores que inciden sobre la demanda futura incluyendo la variabilidad de los mercados individuales y el impacto de las promociones de productos. In addition, U.S. Patent No. 5,459,656 to Fields et al. discloses a system that measures and stores business demand data plus a plurality of intervals of time and projects business demand for products at near future time intervals using demand curves based on percentages. The field system that allows the creation of a plurality of demand curves to determine demand in the near future through the use of defined functions and variables. Projections of business demand for current and near future time intervals can be reviewed in response to variations in the actual data of business demand, which you already received. The fields, however, fail to produce forecasts that proactively take into account many factors that affect future demand including the variability of individual markets and the impact of product promotions.
Por último, la patente de EE.UU.. No. 6.032.125 de Ando describe un sistema de previsión de la demanda que permite a los usuarios pronosticar la demanda con base en algoritmos derivados de datos de diferentes períodos de tiempo, incluyendo los últimos meses, los períodos similares de años anteriores y sus combinaciones. La patente Ando, sin embargo, no tiene en cuenta que los datos sobre la demanda pasada pueden provenir de varias fuentes dentro de una sola organización o cadena de suministro (por ejemplo, los datos de ventas, devoluciones, los datos al por mayor, etc.). Además, Ando no enseña sistemas o métodos para derivar varios modelos de previsión de la demanda y la revisión de esos modelos con el paso del tiempo. Finally, US Patent No. 6,032,125 to Ando describes a demand forecasting system that allows users to forecast demand based on algorithms derived from data from different time periods, including the latest months, similar periods of previous years and their combinations. The Ando patent, however, does not take into account that data on past demand may come from various sources within a single organization or supply chain (for example, sales data, returns, wholesale data, etc. .). In addition, Ando does not teach systems or methods to derive several models of demand forecasting and the revision of those models over time.
Por lo tanto, sigue existiendo una necesidad en la técnica para la mejora de los sistemas y métodos que se pueden desarrollar de manera proactiva modelos alternativos para predecir la demanda a través de múltiples niveles de la cadena de suministro de una organización con el fin de evitar desajustes costosas de la demanda y la oferta.
Therefore, there remains a need in the art for the improvement of systems and methods that alternative models can be proactively developed to predict demand across multiple levels of an organization's supply chain in order to avoid costly mismatches of demand and supply.
DESCRIPCION DETALLADA DE LA INVENCION DETAILED DESCRIPTION OF THE INVENTION
Breve descripción de figuras Brief description of figures
La figura 1 muestra una gráfica de la desviación estándar de un producto. Figure 1 shows a graph of the standard deviation of a product.
La figura 2 representa un diagrama de flujo del proceso de pronostico. Figure 2 represents a flow chart of the forecasting process.
El siguiente método se propone para la predicción de la demanda de productos en establecimiento comercial. The following method is proposed for the prediction of the demand for products in commercial establishment.
Se asume los siguiente: The following are assumed:
Se cuenta con un historial de ventas de al menos 1 año There is a sales history of at least 1 year
Las ventas presentadas representan la demanda total, esto es, el establecimiento no tuvo restricciones de almacenamiento y siempre tuvo disponibilidad de unidades, de modo que las ventas generadas representan la demanda real del producto. The sales presented represent the total demand, that is, the establishment had no storage restrictions and always had unit availability, so that the sales generated represent the actual demand of the product.
Los productos no son estacionales. Es decir, su venta no varía sustancialmente en una época determinada, en ciertos días de la semana o en determinado clima. The products are not seasonal. That is, its sale does not vary substantially at a certain time, on certain days of the week or in a certain climate.
Los productos no son perecederos, de modo que su frescura no impacta en su consumo. The products are not perishable, so that their freshness does not impact their consumption.
Utilización de series de tiempo para estimación de demanda de productos: Use of time series to estimate demand for products:
1. Matriz de correlación 1. Correlation Matrix
El primer paso es calcular la correlación entre productos. En caso de existir uno o varios productos con una correlación por encima de un límite establecido a criterio (por ejemplo, .9) se establece una tabla de correlación entre ambos productos. Ejemplo: The first step is to calculate the correlation between products. If there is one or more products with a correlation above a limit established at the discretion (for example, .9), a correlation table is established between both products. Example:
La tabla mantendrá un registro de los productos cuyo comportamiento se observa con una correlación por encima de un criterio específico (en este caso, > 0.9). En este caso, el producto A y el C tienen una correlación positiva de 0.96, mientras que con el producto B presenta una correlación de 0.90
2. Calcular las estadísticas descriptivas de las ventas y analizar los datos de la primer y última desviación estándar. The table will keep a record of the products whose behavior is observed with a correlation above a specific criterion (in this case,> 0.9). In this case, product A and C have a positive correlation of 0.96, while with product B it has a correlation of 0.90 2. Calculate the descriptive statistics of the sales and analyze the data of the first and last standard deviation.
Se observa la distribución de venta del producto durante el año de historial. Se excluyen de la muestra los valores que caigan en la primera y última desviación estándar. La intención es utilizar los datos para entrenar un modelo de series de tiempo evitando sesgos. The sales distribution of the product is observed during the year of history. Values that fall into the first and last standard deviation are excluded from the sample. The intention is to use the data to train a time series model avoiding bias.
3. Detección de moda en la primera y última desviación estándar. 3. Fashion detection in the first and last standard deviation.
Mediante un análisis de estadísticas descriptivas de las ventas que caen en los datos excluidos en el paso 2 se determina si hay una moda lo suficientemente marcada como para determinar una variación importante en determinados días de la semana o en semanas específicas durante el año. Los datos alojados en este período contendrán las ventas más bajas y las más altas de todo el historial, por lo que es importante considerarlas para la proyección. Through an analysis of descriptive statistics of the sales that fall into the data excluded in step 2, it is determined whether there is a marked enough trend to determine a significant variation on certain days of the week or on specific weeks during the year. The data hosted in this period will contain the lowest and highest sales in the entire history, so it is important to consider them for the projection.
En la gráfica de la figura 1 se aprecia la primera (-3) y última desviación estándar (3) contienen los picos bajos y altos de la demanda de un producto. Si observamos la moda de los días de la semana o de mes con una frecuencia lo suficientemente alta, podemos tomar en cuenta los días específicos para determinar la demanda en un futuro. The graph of Figure 1 shows the first (-3) and last standard deviation (3) containing the low and high peaks of the demand for a product. If we look at the mode of the days of the week or month with a sufficiently high frequency, we can take into account the specific days to determine the demand in the future.
4. HMM con 5 estados ocultos 4. HMM with 5 hidden states
Se utilizan los datos de demanda producto del paso 3 desglosado diariamente para predecir la demanda de los siguientes n días, donde n <= 3 para los productos que requieran compra diaria, y desglosado por semana para los de compra semanal. The demand data from step 3, broken down daily, is used to predict the demand for the next n days, where n <= 3 for products that require daily purchase, and broken down by week for those of weekly purchase.
El modelo oculto de markov se entrena con 5 estados ocultos y se utiliza para predecir la demanda de la siguiente semana o el siguiente día. Esta predicción se considera normal y es la base para la proyección final, que tomará en cuenta los períodos extraordinarios de ventas. The hidden markov model trains with 5 hidden states and is used to predict demand for the following week or the next day. This prediction is considered normal and is the basis for the final projection, which will take into account the extraordinary sales periods.
5. Predicción final 5. Final prediction
Tomando como base la predicción arrojada por la serie de tiempo HMM (X), se consideran los siguientes criterios: Based on the prediction given by the HMM time series (X), the following criteria are considered:
- Si la predicción cae en un período (semana, día del año) en el que se ha detectado una moda en la disminución o aumento porcentual, se toma en cuenta el aumento o distribución y se aplica a la proyección arrojada por el modelo. - If the prediction falls in a period (week, day of the year) in which a mode has been detected in the percentage decrease or increase, the increase or distribution is taken into account and applied to the projection thrown by the model.
- Si el índice de correlación con otro producto es muy alto, se toma en cuenta el aumento o disminución en la predicción de dicho producto (en porcentaje) y se aplica a la proyección arrojada.
- If the correlation index with another product is very high, the increase or decrease in the prediction of said product (in percentage) is taken into account and applied to the projection thrown.
Claims
REIVINDICACIONES
Lo que se reclama es el uso de series de tiempo para la estimación de la demanda de productos utilizando un modelo oculto de markov (Hidden Markov Model) como sigue: i. Recolectar información sobre la venta de los productos en al menos un año. ii. Calcular las estadísticas descriptivas de las ventas y analizar los datos de la primer y última desviación estándar. What is claimed is the use of time series for estimating the demand for products using a hidden markov model (Hidden Markov Model) as follows: i. Collect information on the sale of products in at least one year. ii. Calculate descriptive sales statistics and analyze the data of the first and last standard deviation.
iii. Encontrar si hay una moda lo suficientemente fuerte como para indicar la influencia de las ventas de los productos en días de la semana específicos o en fechas determinadas. iii. Find if there is a fashion strong enough to indicate the influence of product sales on specific days of the week or on specific dates.
iv. Predecir la demanda utilizando un modelo oculto de Markov utilizando como datos de entrenamiento sólo los datos de las dos desviaciones estándar centrales para evitar sesgos. iv. Predict demand using a hidden Markov model using as training data only the data of the two central standard deviations to avoid bias.
v. Complementar la predicción del modelo HMM con 5 estados ocultos con los datos que arroja el análisis de la moda en los extremos de la distribución. La decisión final será como siempre, la del cliente, sin embargo tendrá una fuerte base para determinar si altera la cantidad sugerida o aplica otro criterio de acuerdo a la experiencia. v. Complement the prediction of the HMM model with 5 hidden states with the data shown by the analysis of fashion at the extremes of the distribution. The final decision will be as always, that of the client, however it will have a strong basis to determine if it alters the suggested amount or applies another criterion according to the experience.
El método de la reivindicación anterior contempla la observación de relación entre productos para la predicción de la demanda y detecta la tendencia cíclica en caso de aplicar, es decir, una alza imperceptible para el dependiente en ciertos días de la semana.
The method of the preceding claim contemplates the observation of the relationship between products for the prediction of the demand and detects the cyclical tendency in case of applying, that is, an imperceptible rise for the dependent on certain days of the week.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| PCT/MX2015/000015 WO2016122292A1 (en) | 2015-01-27 | 2015-01-27 | Method for forecasting the demand of non-seasonal and non-perishable products |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| PCT/MX2015/000015 WO2016122292A1 (en) | 2015-01-27 | 2015-01-27 | Method for forecasting the demand of non-seasonal and non-perishable products |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2016122292A1 true WO2016122292A1 (en) | 2016-08-04 |
Family
ID=56543810
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/MX2015/000015 WO2016122292A1 (en) | 2015-01-27 | 2015-01-27 | Method for forecasting the demand of non-seasonal and non-perishable products |
Country Status (1)
| Country | Link |
|---|---|
| WO (1) | WO2016122292A1 (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20020174005A1 (en) * | 2001-05-16 | 2002-11-21 | Perot Systems Corporation | Method and system for assessing and planning business operations |
| US20050234718A1 (en) * | 2004-04-15 | 2005-10-20 | Khimetrics, Inc. | System and method for modeling non-stationary time series using a non-parametric demand profile |
| WO2013175418A1 (en) * | 2012-05-22 | 2013-11-28 | Mobiag, Lda. | System for making available for hire vehicles from a fleet aggregated from a plurality of vehicle fleets |
-
2015
- 2015-01-27 WO PCT/MX2015/000015 patent/WO2016122292A1/en active Application Filing
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20020174005A1 (en) * | 2001-05-16 | 2002-11-21 | Perot Systems Corporation | Method and system for assessing and planning business operations |
| US20050234718A1 (en) * | 2004-04-15 | 2005-10-20 | Khimetrics, Inc. | System and method for modeling non-stationary time series using a non-parametric demand profile |
| WO2013175418A1 (en) * | 2012-05-22 | 2013-11-28 | Mobiag, Lda. | System for making available for hire vehicles from a fleet aggregated from a plurality of vehicle fleets |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Syntetos et al. | Forecasting intermittent inventory demands: simple parametric methods vs. bootstrapping | |
| CN111133460B (en) | Optimization of demand prediction parameters | |
| CN110555578B (en) | Sales prediction method and device | |
| WO2015041778A1 (en) | Product promotion optimization system | |
| Hasin et al. | An ANN approach to demand forecasting in retail trade in Bangladesh | |
| WO2020024718A1 (en) | Method and device for predicting foreign exchange transaction volume | |
| US20240086726A1 (en) | Systems and methods for big data analytics | |
| Aithal | Impact of Domestic, Foreign, and Global Environments on International Business Decisions of Multinational Firms: A Systematic Study | |
| CN115358651A (en) | Method, system and computer readable storage medium for automatic replenishment of retail store | |
| Tadayonrad et al. | A new key performance indicator model for demand forecasting in inventory management considering supply chain reliability and seasonality | |
| Tanaka | Firm heterogeneity and FDI in distribution services | |
| Framinan et al. | Available-to-promise systems in the semiconductor industry: A review of contributions and a preliminary experiment | |
| JP5819363B2 (en) | Demand prediction apparatus and program | |
| WO2016122292A1 (en) | Method for forecasting the demand of non-seasonal and non-perishable products | |
| Chase Jr | Innovations in business forecasting | |
| van Leeuwen et al. | Demand forecasting in hospitality using smoothed demand curves | |
| US20170103433A1 (en) | Generating and Using Estimated Asset Monetary Values and Estimated Macroeconomic Measure Data Values | |
| Sadiakhmatov et al. | Specificity of undivided time series forecasting described with innovation curves | |
| Moreno et al. | The macro-finance environment and asset allocation: A simultaneous equation approach | |
| Akhmetova et al. | The Calculation Methodology of the Synergistic Effect of Digitalization of Retail Services | |
| Karan et al. | The impact of training data tailoring on demand forecasting models in retail | |
| Subramani | Building a Strategic Framework for Retail Supply Chain Analytics | |
| Pahadi et al. | Retail store sale prediction | |
| Huang et al. | New-Product Demand Forecasting for Long-Lived Products | |
| Rajurkar et al. | Development of an integrated modelling framework for retailers of perishable products: a dynamic programming approach |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 15880308 Country of ref document: EP Kind code of ref document: A1 |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| 32PN | Ep: public notification in the ep bulletin as address of the adressee cannot be established |
Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205 DATED 04/01/2018) |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 15880308 Country of ref document: EP Kind code of ref document: A1 |