CROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of and priority to US Provisional Patent Application No. 63/370,047, filed on August 1, 2022, the entire contents of which is hereby incorporated by reference herein.

BACKGROUND

[0002] Embodiments of the present disclosure relate to Supply Chain Management (SCM), and more specifically, to an Al-based supply chain replanning solution. There are many parts of supply chain networks that can introduce uncertainty into a company's strategic planning. The uncertainty is centered on the inability, or difficulties, in accurately and dynamically predicting future demand. Therefore, there is a need for strategic planning to reduce a company's costs, to improve its profits, to avoid costly shortages or periods of excess inventory, as well as to controlling production cycles of products.

BRIEF SUMMARY

[0003] Artificial Intelligence replanning, (Al rePlan) provides a solution that is an automated Al-based optimized planning, controlling, and replanning of supply chain and demand management by multiple scenarios analysis and optimization. According to some embodiments of the present disclosure, methods of and computer program products for planning supply chain activities are provided. In various embodiments, a method for planning supply chain activities is provided. Historical supply chain data from sources internal and external to a company are processed to produce a set of supply chain data features. A predicted demand signal is produced using a first machine learning model trained with the set of supply chain data features. A predicted supply signal is produced using a

1

SUBSTITUTE SHEET ( RULE 26) second machine learning model trained with the set of supply chain data features. A plurality of scenarios is determined using a third machine learning model, wherein the predicted demand signal and the predicted supply signal are inputs to the third machine learning model. The plurality of scenarios are filtered based on attributes of the company.

[0004] In various embodiments, a system is provided including a computing node comprising a computer readable storage medium having program instructions embodied therewith. The program instructions are executable by a processor of the computing node to cause the processor to perform a method. Historical supply chain data from sources internal and external to a company are processed to produce a set of supply chain data features. A predicted demand signal is produced using a first machine learning model trained with the set of supply chain data features. A predicted supply signal is produced using a second machine learning model trained with the set of supply chain data features. A plurality of scenarios is determined using a third machine learning model, wherein the predicted demand signal and the predicted supply signal are inputs to the third machine learning model. The plurality of scenarios are filtered based on attributes of the company.

[0005] In various embodiments, a computer program product for planning supply chain activities is provided including a computer readable storage medium having program instructions embodied therewith. The program instructions are executable by a processor to cause the processor to perform a method. Historical supply chain data from sources internal and external to a company is processed to produce a set of supply chain data features. A predicted demand signal is produced using a first machine learning model trained with the set of supply chain data features. A predicted supply signal is produced using a second machine learning model trained with the set of supply chain data features. A plurality of scenarios is determined using a third machine learning model, wherein the predicted demand signal and the predicted supply signal are inputs to the third machine learning model. The plurality of scenarios are filtered based on attributes of the company.

BRIEF DESCRIPTION OF THE DRAWINGS

[0006] Fig. 1 illustrates a workflow for supply chain and demand planning, replanning, and control according to various embodiments of the present disclosure.

[0007] Fig. 2 illustrates a high-level workflow for supply chain and demand planning, replanning, and control according to various embodiments of the present disclosure.

[0008] Fig. 3 is a flow diagram of example process for planning supply chain activities according to various embodiments of the present disclosure.

[0009] Fig. 4 depicts a computing node according to various embodiments of the present disclosure.

DETAILED DESCRIPTION

[0010] Supply Chain Management (SCM) is the centralized management of the flow process, steps, and requirements that are used to transform raw materials into final products, aiming to gain competitive advantages and value creation for customers. Among all the parts of SCM, one of the more complex is strategic planning because it represents an effort by the suppliers to develop and implement the most efficient and economical planning possible. For this optimization, the manager within a company must look at the company’s logistics and the inventory to minimize shortages and overstocks, to keep costs and inventory levels low.

[0011] A supply chain is responsible for organizing all pieces of this flow, including individuals, companies, activities, resources, technologies, deliveries, and logistics with the goal of getting a product or service to a customer. Supply chains are networks of customers, distributors, and suppliers that are dynamic, coordinated systems involving moving, transforming, transporting, and distributing materials or services. Thus, supply chains as described herein may be interchangeably referred to as supply chain networks. Various embodiments and examples, as used herein, may be provided in the context of a company or organization trying to optimize or strategically plan their supply chain networks.

[0012] A business problem is that there are many parts of supply chain networks that can introduce uncertainty into a company's strategic planning. The uncertainty may be centered on the inability or difficulties in accurately and dynamically predicting future demand within the supply chain network. Therefore, strategic planning may be necessary to reduce a company's costs, to improve its profits, to avoid costly shortages or periods of excess inventory, as well as to control production cycles of products. Overall strategic and accurate planning for the supply chain network offers many opportunities for companies to improve their Key Performance Indicators (KPIs) and operations.

[0013] Today, supply chains are not simply selling and delivering products. Supply chains should provide agility in system integrations, visibility into all the moving parts, and data- driven strategic decision-making in all functional areas of the company. In a dynamic environment with competitors offering similar products at equivalent prices, the agility factor and inventory optimization may be decisive for success or failure of the company. With globalization, supply chains have be more complex and also more sensitive to errors. The ripple effect of errors up and down a supply chain can negatively impact suppliers, distributors, and customers. These suppliers, distributors, and customers can be impacted in a relatively short time frame, for example, in days rather than months. Therefore, companies need the ability to react immediately to changes that will impact their operations.

[0014] In conventional supply chain management solutions available on the market, the majority of supply chain decisions rely on human guesswork, historical scenarios, and other traditional methods of decision-making. However, humans are unable to process all the information and data available to make truly informed decisions in the time frames typically allowed in today’s fast-moving environment. As a result, errors occur and these errors result in order delays, stockouts, low efficiency, and increased costs. These issues caused by the errors may culminate in poor customer experiences and lost revenue. In addition, conventional supply chain management solutions may be costly, such as multiple millions of dollars. The conventional supply chain management solutions also include point solutions that target one specific area of a supply chain such as logistics or production scheduling. The conventional solutions may not be customizable to a customer’s KPI priorities. Conventional solutions may require long implementation times and the replacement of existing infrastructure and workflows. As a result, with the use of conventional solutions, this creates the need for change management programs and re-training of the workforce to learn new tools and processes/workflows. In particular, the conventional solutions do not provide AI- based solutions for planning, replanning, and manipulating supply and demand using continuous scenario simulations. The conventional solutions also do not provide Al-based solutions that include Al -based recommendations for optimization based on a given company’s KPIs, such as profit margin, inventory stock, costs, and more.

[0015] Thus, there is a need for an Al-based supply chain replanning solution, such as Al rePlan, that can work alongside existing workflows and optimize customer-specific company KPIs. In addition, such a solution should be able to analyze a large number of data points, such as millions of data points, to evaluate all possible scenarios in near real-time, and/or continuously learn from the predictions and decisions made. Such a solution may be extremely valuable to any company with a dynamic supply chain.

[0016] Artificial Intelligence replanning, (Al rePlan), is a solution that provides for an automated Al-based optimized plan, control, and replan of a supply chain network as well as demand or supply plan management. The Al rePlan solution may use multiple scenario analysis and scenario optimization. The solution may consider changes and errors in an existing plan over time. The solution may also adapt faster for more accurate replanning and adjusting of supply chain networks in order to optimize company KPIs for multiple scenarios. [0017] Scenario simulation, such as that performed by the Al rePlan solution involves filtering and optimizing various combinations of decisions that could be made along the supply and demand continuum, such as those decisions involving the purchase of items, item inventory, lead time, and the like. The outcome of each scenario matched together with the desired outcome of the company’s KPIs may lead to intelligent decision-making for any supply chain in any particular unit of time, such as a day. An example of different scenarios may be, for example, if a customer needs 100 units of a product, it can order or reorder the units from one supplier, different suppliers, a combination of suppliers, or even order from another distribution center. The Al rePlan solution may consider many or every possible scenario. The solution may choose and present to the company or company management the most advantageous scenario for the company implementing the Al recommended rePlan solution.

[0018] Al rePlan is an automated solution framework using Al that does planning, controlling, and replanning of the supply chain workflow and demand management activities. This solution uses continual analysis of multiple scenarios, identifies deviations from current plans using Al, and provides recommended actions to company decision-makers so they can maximize company KPIs. The solution may be used to assist such decision-makers in selecting the best scenario and making the best decisions at the moment to plan their purchases, move inventory, and adjust future plans. The Al model(s) employed by the solution are built to continually optimize profits, lead times, inventory, and the like. As used herein Al model(s) may be used generally to refer to any machine learning model(s) operating on component(s) of one or more computing nodes. For example, the machine learning model(s) may implement supervised learning algorithms, unsupervised learning algorithms, and/or reinforcement learning algorithms. Non-limiting examples of algorithms used by such machine learning model(s) may include linear or logistic regression, decision tree, random forest, k-nearest neighbor (KNN), logistic regression, apriori algorithms, k- means algorithms, support vector machines (SVM), naive Bayes, dimensionality reduction algorithms, gradient boosting algorithms, and/or the like. An advantage of the Al rePlan solution over conventional solutions is that it is built to deliver results in less time, such as a few minutes or a few hours, with fewer errors, when compared to conventional solutions. In particular, the Al rePlan solution provides speed and accuracy to and remove guesswork from a process that conventionally can take a long time, such days, weeks, or months. The Al rePlan solution provides optimization of supply & demand planning and management, often referred to as Sales & Operational Planning (S&OP).

[0019] Although the description herein uses the terms Al model(s) and Al method(s), as used herein, these term may be used interchangeably with machine learning model(s) and/or machine learning method(s). In addition, although particular such model(s) and method(s) described herein, other known machine learning model(s) and method(s) may be used without departing from the scope and spirit of what is described herein.

[0020] Fig. 1 illustrates a workflow 100 for supply chain and demand planning, replanning, and control according to various embodiments of the present disclosure. One or more aspects of workflow 100 may be performed, by way of example, by one or more computer system/server in a computing node and/or by one or more processors as described herein. Workflow 100 includes internal source 150 and external source 152, which each store data and/or provide the data to the data integration component 110. Internal source 150 may store and/or provide internal data, such as company history data. External source 152 may store and/or provide external data, such as weather, economy, politics, holidays and special events, and competitor information. Supply and demand realized storage/repository 160 may store and/or provide historical data regarding past decisions made by company decision-makers regarding supply and demand in the supply chain network.

[0021] Data integration component 110 receives input from internal source 150, external source 152, and supply and demand realized storage/repository 160. Data integration component 110 integrates the data it receives. Data integration component 110 may combine the internal and external data as well as other data it receives from different sources and unify the data. In particular, data integration component 110 may use signal processing techniques, statistical techniques, data matching techniques, and/or the like to integrate the data. Data integration component 110 may also clean the data by fixing the data, removing incorrect, incorrectly formatted, duplicate, or incomplete data among the data that it receives. The output of data integration component 110 can be provided to an integrated and clean data storage/repository 154. The integration and cleaning of data performed by data integration component 110 may be used to understand historical data and to understand the behavior of sales, distributors and supply, lead time, stock, company action, and business rules in a unified manner. Such data and insights may be used to train Al, statistical, and/or math model(s).

[0022] Data feature extraction component 112 may receive and/or access data, such as the integrated and clean data output by the data integration component 110 and stored in data storage/repository 154. Data feature extraction component 112 may extract as much important information as possible from the data it receives. For example, data feature extraction component 112 may extract from the data patterns, features, statistical samples, and/or correlations. Data feature extraction component 112 may then aggregate and group this data. The output of data feature extraction component 112 may be provided to an analytical and aggregated data with business features storage/repository 156. The data in the analytical and aggregated data with business features storage/repository 156 may be provided to both the demand forecasting component 114 and supplier behavior component 120.

[0023] Demand forecasting component 114 may receive and/or access data, output by data feature extraction component 112 and stored in the analytical and aggregated data with business features storage/repository 156. For example, demand forecasting component 114 may receive historical sales behavior, inventory information, supplier lead time, and/or external data, such as weather data, economic data, political data, data regarding holidays and/or special events, and/or competitor information, and may include aggregated patterns, features, statistical samples, or correlations of such data.

[0024] In various embodiments, using the data it receives and/or accesses, the demand forecasting component 114 may train various demand Artificial Intelligence (Al) model(s) and methods in order to predict/forecast demand behavior. The Al model(s) may use traditional methods for regression, such as signal point and distributional forecasting, time series forecasting, and/or finite element machine regression methods. The Al model(s) may be trained to be able to predict demand results by granularity. Examples of such granularity may include demand at a distribution center, at a department, for a product group or SKU, and/or total monthly demand. The demand forecasting component 114 may output its result, such as a trained or partially trained demand Al model(s), to demand opening component 116

[0025] In various embodiments, using the data it receives and/or accesses, the demand forecasting component 114 may output a preliminary predicted demand signal to demand opening component 116. This signal may be generated by processing the received data using the trained or partially trained demand Al model(s). The preliminary predicted demand signal may predict demand results by granularity and may be output to demand opening component 116. [0026] Demand opening component 116 may receive the output(s) of the demand forecasting component 114 and may perform Al methods of opening the data and/or apply Al model(s) to the output that it receives. In particular, these Al methods of opening may include granularity adjustments to the data and/or models output by the demand forecasting component 114. For example, the granularity adjustments may include opening monthly demand per day data with confidence level(s) for this data.

[0027] In various embodiments, demand opening component 116 may further train the demand Al model(s) trained at demand forecasting component 114. The demand opening component 116 may output its result, such as a trained or partially trained demand Al model(s), to demand sensing component 118. The output of the demand opening component 116, such as the trained or partially trained demand Al model(s), may be used to output a preliminary predicted demand signal and/or to plan the demand in the supply chain network. [0028] In various embodiments, using the preliminary predicted demand signal and/or other data it receives, the demand opening component 116 may output, to demand sensing component 118, another preliminary predicted demand signal. This signal may be generated by processing the received data using the trained or partially trained demand Al model(s). Such a preliminary predicted demand signal may be continuous and smooth during the course of each day of temporal granularity. Such a preliminary predicted demand signal may allow for the more accurate prediction of the demand. Such a signal may also produce a lower level of uncertainty in the supply chain network.

[0029] Demand sensing component 118 may receive the output(s) of the demand opening component and may perform demand signal correction within the demand Al model(s) and/or the preliminary predicted demand signals in order to decrease the difference between a predicted demand signal and an actual value of the demand signal. In particular, the outputs of the demand Al model(s) produced after the demand forecasting component 114 and/or the demand opening component 116 operate may be adjusted over time. For example, the demand Al model(s) may output a preliminary expected prediction for the demand signal. An actual value of the demand signal may be known, for example, based on past data regarding actual demand. The difference between the preliminary expected prediction for the demand signal and the actual value of the demand signal may be determined. This difference may be an absolute error value. Demand sensing component 118 may compute this absolute error value. Demand sensing component 118 may minimize this absolute error value by adjusting the demand Al model(s) and/or the preliminary expected prediction for the demand signal as needed.

[0030] Thus, in the short, the medium, and the long term, even if the demand Al model(s) start making mistakes frequently, the error associated with the mistakes may be minimized over time as the adjustments occur. Such demand signal corrections may decrease the difference between the predicted value and the actual value of the demand signal. Correction signal technique(s), such as spread, adjusted, and re-computation signal technique(s), may be used for the adjustment of the demand Al model(s) and/or the expected prediction for the demand signal. The adjustment may be based on correlations of previous predictions.

[0031] In various embodiments, demand sensing component 118 may further train the demand Al model(s) trained at demand opening component 116 as described above. The demand sensing component 118 may output its result, such as adjusted trained or partially trained demand Al model(s), to supply and demand Al integration component 124. The output of the demand sensing component 118, such as the trained or partially trained demand Al model(s), may be used to output a predicted demand signal and/or to plan the demand in the supply chain network.

[0032] In various embodiments, using the preliminary predicted demand signal and/or other data it receives, the demand sensing component 118 may output to Al integration component 124 a predicted demand signal and associated confidence intervals. This signal and confidence intervals may be generated by processing the received data using the trained or partially trained demand Al model(s). In particular, the trained or partially trained demand Al model(s) may output the predicted demand signal and associated confidence intervals to replan the demand in the supply chain network.

[0033] Within the demand forecasting component 114, demand opening component 116, and/or demand sensing component 118 the demand Al model(s) and/or Al method(s) being used may include traditional methods for regression, such as single point and distributional forecasting, time series forecasting, and finite element machine regression methods. The inputs of the Al model(s) and Al method(s), discussed above, may include historical sales behavior, inventory information, suppliers lead time, external data, such as weather, economy, politics, holidays and special events, and competitor information, and/or the like. For example, a single demand Al model that that makes use of these techniques may be trained and/or retrained in near-real-time (e.g., many times a day) using the components 114, 116, and/or 118. The frequency of training and/or retraining may be selectable by a user, selected automatically, and/or dependent on the company in which the Al rePlan solution is deployed. As described above, a predicted demand signal with associated confidence intervals may be generated using the trained or partially trained demand Al model(s). In particular, the output predicted demand signal may include information and/or attributes regarding the demand at varying granularities such as date, distributor center/region of the demand, SKU(s), quantity lower bound(s), quantity upper bound(s), forecast quantities, and/or the like.

[0034] Supplier behavior component 120 may receive and/or access data, output by data feature extraction component 112 and stored in the analytical and aggregated data with business features storage/repository 156. For example, supplier behavior component 120 may receive historical sales behavior, inventory information, supplier lead time, and/or external data, such as weather data, economic data, political data, data regarding holidays and/or special events, and/or competitor information, and may include aggregated patterns, features, statistical samples, or correlations of such data.

[0035] In various embodiments, using the data it receives and/or accesses, the supplier behavior component 120 may train various supply Al model(s) and methods in order to predict/forecast supplier and supply behavior. The Al model(s) may use traditional methods for regression, such as signal point and distributional forecasting, time series forecasting, finite element machine regression methods, mathematical techniques, and/or statistical modeling, such as trends detection, pattern recognition, use of confidence intervals, use of correlation matrices, and/or the like. The Al model(s) may be used to identify and predict estimated delivery times/dates by producing a predicted supply signal with confidence intervals. In some embodiments, a final shipment schedule may be predicted. The supplier behavior component 120 may output its result, such as a trained or partially trained supply Al model(s), to supply chain sensing component 122.

[0036] In various embodiments, using the data it receives and/or accesses, the supplier behavior component 120 may output a preliminary predicted supply signal that includes an expected delivery time and quantity per supplier, with the confidence intervals for each time. This signal and the confidence intervals may be generated by processing the received and/or accessed data using the trained or partially trained supply Al model(s). For example, a supplier A may have a 99% confidence level that it will deliver the purchase order in 7 days. In addition, the supplier A may have a 98% confidence level that it will deliver a quantity in the range [950; 1050] items considering the 1,000 units ordered. The supplier behavior component 120 may output the predicted supply signal and confidence intervals to supply chain sensing component 122, and this signal and confidence intervals may be used to plan the suppliers’ deliveries in the supply chain network.

[0037] Supply chain sensing component 122 may receive the output(s) of the supplier behavior component 120 and may perform supply signal correction in order to decrease the difference between the preliminary predicted supply signal and an actual value of the supply signal. In particular, the outputs of the supply Al model(s) produced after the supplier behavior component 120 operates may be adjusted over time. For example, these model(s) may output a preliminary expected prediction for the supply signal and its associated confidence intervals. An actual value of the supply signal may be known, for example, based on past data regarding actual supply levels. The difference between the preliminary expected prediction for the supply signal and the actual value of the supply signal may be determined. This difference may be an absolute error value. Supply chain sensing component 122 may compute this absolute error value. Supply chain sensing component 122 may minimize this absolute error value by adjusting the supply Al model(s) and/or the preliminary expected prediction for the supply signal as needed.

[0038] Thus, in the short, medium, and long term, even if the model(s) start making mistakes frequently, the error associated with the mistakes may be minimized over time as the adjustments occur. Correction signal technique(s), such as spread, adjusted, and recomputation signal technique(s), may be used for the adjustment of the supply Al model(s) and/or the expected prediction for the supply signal. The adjustment may be based on correlations of previous predictions. Such supply signal corrections may decrease the difference between the predicted value and the actual value of the supply signal and may also produce more accurate confidence intervals.

[0039] In various embodiments, supply chain sensing component 122 may further train the supply Al model(s) trained at supplier behavior component 120 as described above. The supply chain sensing component 122 may output its result, such as adjusted trained or partially trained supply Al model(s), to supply and demand Al integration component 124. The output of the supply chain sensing component 122, such as the trained or partially trained supply Al model(s), may be used to output a predicted supply signal and associated confidence intervals and/or to replan the suppliers’ deliveries in the supply chain network. [0040] In various embodiments, using the preliminary predicted demand signal and/or other data it receives, the supply chain sensing component 122 may output to Al integration component 124 predicted supply signal and its associated confidence intervals. This signal and confidence intervals may be generated by processing the received data using the trained or partially trained supply Al model(s). In particular, the trained or partially trained demand Al model(s) may output the predicted supply signal and associated confidence intervals to replan the supply in the supply chain network.

[0041] Supply chain sensing component 122 may use business rules and previous output to manage supply and stock in the supply chain network. Using the output of the aforementioned supply Al model(s), decisions regarding supply and stock may be made in near real-time. In particular, these decisions may be made based on such model(s) to adjust orders and avoid two problems: stockout and overstock. Thus, for example, in situations of high demand that would be centralized in one supplier which has a high probability of generating disruption, one solution may be to statistically distribute the order to different suppliers or different distribution centers, thereby reducing the costs for the company attempting to optimize its supply chain network. Another solution may involve the use of the sensitivity of delayed information. Thus, advanced adjustments may be made to accelerate business decisions. As discussed above, these supply predictions/forecasts may be adjusted over time so that error is minimized. [0042] Within the supplier behavior component 120 and/or supply chain sensing component 122 the statistical, math, and supply Al model(s), Al method(s), and/or the statistical distribution model(s) being used may include traditional methods for regression, such as single point and distributional forecasting, time series forecasting, finite element machine regression methods, and statistical modelling, such as trends detection, pattern recognition, confidence intervals, correlation matrix, etc., and the like. The inputs of the statistical, math, and Al model(s), Al method(s), and/or the statistical distribution models, discussed above, may include historical sales behavior, inventory information, suppliers lead time, and external data, such as weather, economy, politics, holidays and special events, and competitor information, and/or the like. Past data as well as current and/or past business rules may also be input to the model(s). For example, a single supply Al model that that makes use of these techniques may be trained and/or retrained in near-real-time (e.g., many times a day) using the components 120 and 122. The frequency of training and/or retraining may be selectable by a user, selected automatically, and/or dependent on the company in which the Al rePlan solution is deployed. A predicted supply signal with associated confidence intervals may be generated using the trained supply Al model. In particular, the output supply signal may include information and/or attributes regarding the supply at varying granularities such as expected delivery time/date, geographies/locations of the suppliers, quantity per supplier, and confidence/probability of the supplier’s successful delivery, and/or the like.

[0043] Supply and demand Al integration component 124 may receive the outputs from the demand sensing component 118 and from the supply chain sensing component 122, and may integrate these outputs. In various embodiments, supply and demand Al integration component 124 may receive the adjusted trained or partially trained demand Al model(s) and the adjusted trained or partially trained supply Al model(s). The supply and demand Al integration component 124 may perform integration of the model(s) and/or the signals that it receives. In particular, the adjusted trained or partially trained demand Al model(s) and the trained or partially trained supply Al model(s) may be combined into an integrated supply and demand Al model(s).

[0044] In various embodiments, the supply and demand Al integration component 124 may receive the outputs of the demand Al model(s) and the supply Al model(s), such as a predicted demand signal and a predicted supply signal and their associated confidence intervals, if any. In various embodiments, the supply and demand Al integration component 124 may receive business rules associated with the company. The supply and demand Al integration component 124 may perform integration of the signals that it receives. In particular, the predicted demand signal with confidence intervals and the predicted supply signal with confidence intervals may be combined into an integrated supply and demand signals.

[0045] In various embodiments, Al integration component 124 may make use of the aforementioned integrated supply and demand Al model(s) to process the signals that it receives. In various embodiments, Al integration component 124 may make use of Al model(s) based on math and statistical modelling, such as math formulas, trends detection, pattern recognition, confidence intervals, correlation matrix, etc., bio-inspired optimization methods, such as particle swarm optimization, genetic algorithms, social spider optimization, etc., and/or the like to process the signals that it receives. After processing the signals that it receives, Al integration component 124 may output a large number of scenarios, each relating to the demand and the supply in the supply chain network, and confidence value(s) for each scenario.

[0046] For example, if demand signal indicates that the demand has increased with a sufficiently high confidence, an adjustment to the supply signal should be made to adjust the suppliers’ deliveries. As another example, if the signals indicate that the stock is low with a sufficiently high confidence but the demand has decreased with a sufficiently high confidence, it may not be necessary to place further purchase orders. Thus, the data regarding the supply and the demand in the supply chain network may be analyzed together in order to allow the Al replanning to optimally interconnect them using the supply and demand Al integration component 124. The integration of the supply and demand signals in this way may work alongside existing workflows. As discussed above, the output of the supply and demand Al integration component 124 may be a large number of scenarios and confidence value(s) for each scenario. As an example, a scenario may be that there is a demand for 1,000 units at the end of a month and supplier A may have a 95% confidence level that it will deliver a quantity of 500 units and supplier B may have a 80% confidence level that it will deliver a quantity of 500 units. The output of supply and demand Al integration component 124 may be provided to scenario filtering component 126.

[0047] Scenario filtering component 126 may receive the output of the supply and demand Al integration component 124. In various embodiments, scenario filtering component 126 may also receive statistical model(s) relating the scenarios, business rules, business requirements, and or other information relevant to filtering the scenarios it receives. Many possible scenarios may be generated when the supply and demand signals are linked by the supply and demand Al integration component 124. It may be infeasible to humanly analyze all of the large number of scenarios output by the supply and demand Al integration component 124. However, many scenarios, and at times most scenarios, may be statistically irrelevant due to low confidence value(s) associated with the scenarios. Many scenarios, and at times most scenarios, may be irrelevant to the business because they may be impractical to company logistics or would result in issues regarding a low profit margin and/or excessive company costs. Scenario filtering component 126 may filter out many scenarios that it receives and may provide as output the scenarios that may be likely to occur and/or only the scenarios that may be likely to occur. Scenario filtering component 126 may also eliminate scenarios that are outliers and those considered “bad” scenarios, such as those that may cause any known problems or issues within the company. In various embodiments, in order to filter out scenarios, scenario filtering component 126 may use statistical model(s) relating the scenarios, business rules, business requirements, and or other information relevant to filtering the scenarios. Scenario filtering component 126 may output feasible scenarios that may be likely to occur to scenario optimization component 128. The data output by scenario filtering component 126 may be used to plan and/or replan the supply and demand in the supply chain network based on feasible scenarios.

[0048] Scenario optimization component 128 may receive the output of scenario filtering component 126. In various embodiments, scenario optimization component 128 may receive multiple feasible scenarios from scenario filtering component 126. These received scenarios may be optimized by filtering the scenarios based on company KPIs. For example, these scenarios may be filtered based on KPIs such as profit margin, stock inventory, lead time, relevant purchase orders, logistic time, company costs, or the like. Thus, the received multiple feasible scenarios may be filtered by scenario optimization component 128 to produce an optimal set of scenarios that meet or exceed company KPIs. Scenario optimization component 129 may output these optimal set of scenarios to monitor analysis component 130. These optimal set of scenarios, which may be output by scenario optimization component 128, may be used to plan and /or replan the supply and demand in the supply chain network.

[0049] Within the supply and demand Al integration component 124, scenario filtering component 126, and/or scenario optimization component 128 the supply and demand integration Al model(s) and filtering method(s) being used may include math and statistical modelling, such as math formulas, trends detection, pattern recognition, confidence intervals, correlation matrices, etc., bio-inspired optimization methods, such as particle swarm optimization, genetic algorithms, social spider optimization, etc., and/or the like. The inputs and outputs of these supply and demand integration Al model(s) and filtering method(s) may be as described above, with reference to components 124, 126, and 128.

[0050] Monitor analysis component 130 may receive the output of scenario optimization component 128. In various embodiments, monitor analysis component 130 may condense all of the set of optimized scenarios generated as output by scenario optimization component 128. In particular monitor analysis component 130 may transform the set of optimized scenarios to a format suitable for business planning of existing workflows, such as customer(s) internal purchase systems. In various embodiments, the resulting newly formatted scenarios 158 may be provided as output to company decision-makers. The company decision-makers may make decisions based on the information and these newly formatted scenarios that they receive and may execute on these decisions by adjusting the supply and demand accordingly. In various embodiments, the decisions, such as purchasing/supplying of goods, may be made automatically by the Al rePlan solution. The formatted set of optimized scenarios and/or the decisions may also be exported and/or stored by supply and demand realized storage/repository 160. Supply and demand realized storage/repository 160 may store and/or provide historical data regarding past decisions made by company decision-makers regarding supply and demand in the supply chain network. The newly formatted scenarios, which may be output by monitor analysis component 130, may be used to plan and /or replan the supply and demand in the supply chain network.

[0051] Although not shown in Fig. 1, analytic data, such as explainable Al analytic data, may be generated and provided at by each of the components 110. . . 130 to all other components, storages/repositories, and decision-makers in Fig. 1. This analytic data may also be provided to knowledge workers. Such analytic data may include, for example, all company KPIs, suggestions for optimal scenarios, the forecasting/impact in the business workflow of each decision, and the like. Where such analytic data is generated based on the use of Al model(s), the data may also include information regarding how and/or why the Al model(s) made decisions and/or output various data.

[0052] Fig. 2 illustrates a high-level workflow 200 for supply chain and demand planning, replanning, and control according to various embodiments of the present disclosure. Workflow 200 may include components that are similar in form and function to the components in workflow 100, described herein. Workflow 200 shows a more condensed version of workflow 100. One or more aspects of workflow 200 may be performed, by way of example, by one or more computer system/server in a computing node and/or by one or more processors as described herein. Workflow 200 may include demand sensing Al rePlan inputs 218 and supply sensing Al rePlan inputs 222. Demand sensing Al rePlan inputs 218 inputs may be the output(s) of demand sensing component 118 and/or supply and demand Al integration component 124. Supply sensing Al rePlan inputs 222 may be the output(s) of supply chain sensing component 122 and/or supply and demand Al integration component 124. Demand sensing Al rePlan inputs 218 and Al rePlan inputs 222 may be provided, as planning inputs, to scenario filtering component 226. Scenario filtering component 226 may be similar in form and function to scenario filtering component 126. Scenario filtering component 226 may receive the same inputs and provide the same outputs as scenario filtering component 126. Scenario filtering component 226 may provide its outputs to scenario optimization component 228. Scenario optimization component 228 may be similar in form and function to scenario optimization component 128. Scenario optimization component 228 may receive the same inputs and provide the same outputs as scenario optimization component 128. Scenario optimization component 228 may provide its outputs to monitor analysis component 230. Monitor analysis component 230 may be similar in form and function to monitor analysis component 130. Monitor analysis component 230 may receive the same inputs and provide the same outputs as monitor analysis component 130. Monitor analysis component 230 may provide its outputs, such as formatted scenarios 158, to company decision-makers. The company decision-makers may make decisions based on the information and these newly formatted scenarios that they receive and may execute on these decisions by authorizing supply and demand plan changes and adjusting the supply and demand accordingly.

[0053] Analytic data 270, such as explainable Al analytic data, may be generated and provided at by each of the components 218, 222, 226, 228, and 230 to all other components. This analytic data may also be provided to knowledge workers. Such analytic data may include, for example, all company KPIs, suggestions for optimal scenarios, the forecasting/impact in the business workflow of each decision, and the like. Where such analytic data is generated based on the use of Al model(s), the data may also include information regarding how and/or why the Al model(s) made decisions and/or output various data.

[0054] Although the description herein uses the terms Al model(s) and Al method(s), as used herein, these term may be used interchangeably with machine learning model(s) and/or machine learning method(s). In addition, although particular such model(s) and method(s) described herein, other known machine learning model(s) and method(s) may be used without departing from the scope and spirit of what is described herein. Additionally, although prediction of the demand and/or supply is described herein, prediction of production can similarly be predicted, and a production plan may be determined, without departing from the scope or spirit of what is described herein. For example, production may be predicted using production history and by determining production signal(s) with confidence interval(s), and by performing production sensing rather than what is described for predicting demand and/or supply as described herein.

[0055] Fig. 3 is a flow diagram of example process for planning supply chain activities 300 according to various embodiments of the present disclosure. The process 300 may be performed, by way of example, by one or more computer system/server in a computing node and/or by one or more processors as described herein. The operations of the process 300 may be performed by any of the components described herein, such as the components described in connection with Figs. 1 or 2. While the operations of the process 300 are described in a particular order, it should be understood that the order may be modified and operations may be performed in parallel. Moreover, it should be understood that operations may be added or omitted.

[0056] At 310 historical supply chain data from sources internal and external to a company may be processed to produce a set of supply chain data features. The historical supply chain data may include historical data from a supply and demand realized storage/repository. For example, the historical supply chain data may be received and processed by data integration component 110 from, for example, internal source 150, external source 152, and/or supply and demand realized storage/repository 160 as described herein.

[0057] At 320 a predicted demand signal may be produced using a first machine learning model trained with the set of supply chain data features. The predicted demand signal may include attributes regarding the demand on a date, demand by a distributor center/region, demanded SKU(s), a lower bound on the quantity of demand, or an upper bound on the quantity of demand. The first machine learning model may be configured to perform regression, time series forecasting, or finite element machine regression. For example, demand forecasting component 114, demand opening component 116, and/or demand sensing component 118 may make use of and/or produce the first machine learning model. In addition, these components and/or the model may produce the predicted demand signal. [0058] At 330 a predicted supply signal may be produced using a second machine learning model trained with the set of supply chain data features. The predicted supply signal may include an expected delivery time and quantity per supplier, with the confidence intervals for each time. The second machine learning model may be configured to perform regression, time series forecasting, finite element machine regression, or statistical modeling. For example, supplier behavior component 120 and/or supply chain sensing component 122 may make use of and/or produce the second machine learning model. In addition, these components and/or the model may produce the predicted supply signal.

[0059] At 340 a plurality of scenarios may be determined using a third machine learning model, wherein the predicted demand signal and the predicted supply signal are inputs to the third machine learning model. The third machine learning model may be configured to perform math and statistical modelling or inspired optimization methods. For example, supply and demand Al integration component 124, scenario filtering component 126, and/or scenario optimization component 128 may make use of and/or produce the third machine learning model. In addition, these components and/or the model may produce the plurality of scenarios.

[0060] At 350 the plurality of scenarios may be filtered based on attributes of the company. The company attributes may include company scenario feasibility information and company key performance indicators. For example, scenario filtering component 126, scenario optimization component 128, and/or monitor analysis component 130 may filter scenarios based on attributes of the company.

[0061] As shown in Fig. 4, computer system/server 12 in computing node 10 is shown in the form of a general -purpose computing device. The components of computer system/server 12 may include, but are not limited to, one or more processors or processing units 16, a system memory 28, and a bus 18 that couples various system components including system memory

28 to processor 16.

[0062] Bus 18 represents one or more of any of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, and a processor or local bus using any of a variety of bus architectures. By way of example, and not limitation, such architectures include Industry Standard Architecture (ISA) bus, Micro Channel Architecture (MCA) bus, Enhanced ISA (EISA) bus, Video Electronics Standards Association (VESA) local bus, Peripheral Component Interconnect (PCI) bus, Peripheral Component Interconnect Express (PCIe), and Advanced Microcontroller Bus Architecture (AMBA).

[0063] Computer system/server 12 typically includes a variety of computer system readable media. Such media may be any available media that is accessible by computer system/server 12, and it includes both volatile and non-volatile media, removable and non-removable media.

[0064] System memory 28 can include computer system readable media in the form of volatile memory, such as random access memory (RAM) 30 and/or cache memory 32. Computer system/server 12 may further include other removable/non-removable, volatile/non-volatile computer system storage media. By way of example only, storage system 34 can be provided for reading from and writing to a non-removable, non-volatile magnetic media (not shown and typically called a "hard drive"). Although not shown, a magnetic disk drive for reading from and writing to a removable, non-volatile magnetic disk (e.g., a "floppy disk"), and an optical disk drive for reading from or writing to a removable, non-volatile optical disk such as a CD-ROM, DVD-ROM or other optical media can be provided. In such instances, each can be connected to bus 18 by one or more data media interfaces. As will be further depicted and described below, memory 28 may include at least one program product having a set (e.g., at least one) of program modules that are configured to carry out the functions of embodiments of the disclosure.

[0065] Program/utility 40, having a set (at least one) of program modules 42, may be stored in memory 28 by way of example, and not limitation, as well as an operating system, one or more application programs, other program modules, and program data. Each of the operating system, one or more application programs, other program modules, and program data or some combination thereof, may include an implementation of a networking environment. Program modules 42 generally carry out the functions and/or methodologies of embodiments as described herein.

[0066] Computer system/server 12 may also communicate with one or more external devices 14 such as a keyboard, a pointing device, a display 24, etc.; one or more devices that enable a user to interact with computer system/server 12; and/or any devices (e.g., network card, modem, etc.) that enable computer system/server 12 to communicate with one or more other computing devices. Such communication can occur via Input/Output (I/O) interfaces 22.

Still yet, computer system/server 12 can communicate with one or more networks such as a local area network (LAN), a general wide area network (WAN), and/or a public network (e.g., the Internet) via network adapter 20. As depicted, network adapter 20 communicates with the other components of computer system/server 12 via bus 18. It should be understood that although not shown, other hardware and/or software components could be used in conjunction with computer system/server 12. Examples include, but are not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data archival storage systems, etc.

[0067] The present disclosure may be embodied as a system, a method, and/or a computer program product. The computer program product may include a computer readable storage medium (or media) having computer readable program instructions thereon for causing a processor to carry out aspects of the present disclosure.

[0068] The computer readable storage medium can be a tangible device that can retain and store instructions for use by an instruction execution device. The computer readable storage medium may be, for example, but is not limited to, an electronic storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. A non-exhaustive list of more specific examples of the computer readable storage medium includes the following: a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), a static random access memory (SRAM), a portable compact disc read-only memory (CD- ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, a mechanically encoded device such as punch-cards or raised structures in a groove having instructions recorded thereon, and any suitable combination of the foregoing. A computer readable storage medium, as used herein, may be signals, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through a waveguide or other transmission media (e.g., light pulses passing through a fiber-optic cable), or electrical signals transmitted through a wire.

[0069] Computer readable program instructions described herein can be downloaded to respective computing/processing devices from a computer readable storage medium or to an external computer or external storage device via a network, for example, the Internet, a local area network, a wide area network and/or a wireless network. The network may comprise copper transmission cables, optical transmission fibers, wireless transmission, routers, firewalls, switches, gateway computers and/or edge servers. A network adapter card or network interface in each computing/processing device receives computer readable program instructions from the network and forwards the computer readable program instructions for storage in a computer readable storage medium within the respective computing/processing device.

[0070] Computer readable program instructions for carrying out operations of the present disclosure may be assembler instructions, instruction-set-architecture (ISA) instructions, machine instructions, machine dependent instructions, microcode, firmware instructions, state-setting data, or either source code or object code written in any combination of one or more programming languages, including an object oriented programming language such as Smalltalk, C++ or the like, and conventional procedural programming languages, such as the “C” programming language or similar programming languages. The computer readable program instructions may execute entirely on the user’s computer, partly on the user’s computer, as a stand-alone software package, partly on the user’s computer and partly on a remote computer or entirely on the remote computer or server. In the latter scenario, the remote computer may be connected to the user’s computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet Service Provider). In some embodiments, electronic circuitry including, for example, programmable logic circuitry, field-programmable gate arrays (FPGA), or programmable logic arrays (PLA) may execute the computer readable program instructions by utilizing state information of the computer readable program instructions to personalize the electronic circuitry, in order to perform aspects of the present disclosure.

[0071] Aspects of the present disclosure are described herein with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the disclosure. It will be understood that each block of the flowchart illustrations and/or block diagrams, and combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer readable program instructions.

[0072] These computer readable program instructions may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions/acts specified in the flowchart and/or block diagram block or blocks. These computer readable program instructions may also be stored in a computer readable storage medium that can direct a computer, a programmable data processing apparatus, and/or other devices to function in a particular manner, such that the computer readable storage medium having instructions stored therein comprises an article of manufacture including instructions which implement aspects of the function/act specified in the flowchart and/or block diagram block or blocks.

[0073] The computer readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable apparatus or other device to produce a computer implemented process, such that the instructions which execute on the computer, other programmable apparatus, or other device implement the functions/acts specified in the flowchart and/or block diagram block or blocks.

[0074] The flowchart and block diagrams in the Figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of instructions, which comprises one or more executable instructions for implementing the specified logical function(s). In some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems that perform the specified functions or acts or carry out combinations of special purpose hardware and computer instructions.

[0075] The descriptions of the various embodiments of the present disclosure have been presented for purposes of illustration, but are not intended to be exhaustive or limited to the embodiments disclosed. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The terminology used herein was chosen to best explain the principles of the embodiments, the practical application or technical improvement over technologies found in the marketplace, or to enable others of ordinary skill in the art to understand the embodiments disclosed herein.