Description:FIELD OF THE INVENTION
[001] The present invention relates to a system for dynamic capacity orchestration in a segmented supply chain to maximize customer experience. Particularly, the present invention relates to a system to provide spillage protection and speed protection to the customers throughout the supply chain process by determining the optimal capacity distribution across the cohorts through a computational model.

BACKGROUND OF THE INVENTION

[002] Supply chain refers to the process of delivering goods or products from the supplier to the end consumer. A well-managed supply chain results into lower costs, greater efficiency in order to ensure that goods are delivered on time and meet customer demands and satisfaction. To achieve an efficient supply chain, it is essential to provide the best service to the customers at every stage, from manufacturing to retail. During sales, there are limited number of employees and large volume of products to sell in supply chain. As a result, there is a requirement for spillage protection and speed protection throughout the entire sales period. However, capacity reservations become difficult during events due to increased overflow or demand surges. Speed protection ensures that the products move through the supply chain at the optimal pace to meet customer demand and business goals.

[003] Current system fails in prioritizing orders of multiple cohorts with optimal resource utilization. The current systems fail in prioritizing custom defined cohorts that are needed to meet business needs.

[004] In addition, the existing systems faces challenges when trying to expand and cater to a dynamic cohort definition. The system may handle the current load for dynamic cohort definition, however, the system faces difficulty in scaling the operation in supply chain to meet the more complex, variable needs of multiple customer segments. The lack of flexibility and efficiency in scaling up to meet dynamic and multiple cohort definition is a core challenge in the system. Therefore, there is a requirement to provide a system to handle various types of cohorts and not single type of cohort.

[005] Prior art US11282022B2 discloses a system predicting a supply chain performance. The system receives supply chain data for delivery of a product. The supply chain data includes input signals comprising operational plans and observed supply chain operational metrics. The input signals include a delivery date of the product. The system automatically generating predicted supply chain operational metrics across including a value at risk that is predicted for the product. The system automatically infers causal factors that impact the predicted supply chain operational metrics including impacting the value at risk that is predicted for the product. The system automatically generates action recommendations for the supply chain. An action recommendation includes a first predicted value impact and a sequence of actions impacting the product the delivery date of the product and the value at risk that is predicted for the product. However, the system fails to segment customers/transactions based on the customer’s attributes in order to offer better speed to the high priority cohort.

[006] Therefore, in view of above mentioned problems, there is a requirement of a system offering better speed in supply chain management for all defined cohorts/customer-segments for a predefined duration through dynamic capacity reallocation.

OBJECTIVE OF THE INVENTION

[007] The primary objective of the present invention is to provide a system for dynamic capacity orchestration in a segmented supply chain to maximize customer experience.

[008] Yet another objective of the present invention is to provide better protection and speed to the multiple cohorts with priority associated with them.

[009] Yet another objective of the present invention is to segment customers/ transactions/ supply chain network based on the customer’s attributes and provide a differentiated service experience to the specific cohorts.

[0010] Yet another objective of the present invention is to employ a computational model for capacity initialization and dynamic capacity re-allocations in the supply chain to prioritize orders from protected cohorts to provide better speed based on customer expectations.

[0011] Yet another objective of the present invention is to ensure flexible utilization of the system during any part of the year based on cohort definitions.

BRIEF DESCRIPTION OF DRAWINGS

[0012] The present invention will be better understood after reading the following detailed description of the presently preferred aspects thereof with reference to the appended drawings, in which the features, other aspects and advantages of certain exemplary embodiments of the invention will be more apparent from the accompanying drawing in which:

[0013] Figure 1 illustrates an exemplary flow chart to demonstrate spillage view in single cohort scenario.

[0014] Figure 2 illustrates an exemplary flowchart of cut-offs and the case of two capacity pools (i.e. Non-Protected and Protected).

[0015] Figure 3 illustrates an exemplary flowchart to demonstrate under-utilization view for non-protected and protected view.

[0016] Figure 4 illustrates the system for predictive unit and reactive unit in capacity allocations.

[0017] Figure 5 illustrates a graphical representation of Demand, Capacity and Spillage for a major event with event days & recovery and at the end of recovery days, spillage is expected to be at BAU level number by end of recovery.

[0018] Figure 6 illustrates a block diagram of simulation engine with inputs & outputs and interaction with three modules Forecasting, Optimization and Guardrail.

[0019] Figure 7 (a) illustrates a graphical representation of comparison of Spillage between FCFS and model results for 4 major FCs/ Warehouses.

[0020] Figure 7 (b) illustrates a graphical representation of comparison of protection provided to VIP cohort between FCFS and Model results for 4 major FCs/Warehouses.

[0021] Figure 7 (c) illustrates a graphical representation of comparison of protection provided to High-ASP cohort between FCFS and Model results for 4 major FCs/Warehouses.

SUMMARY OF THE INVENTION

[0022] The present invention relates to a system for predictive interventions in capacity allocations to prevent spillage and under-utilization of the cohorts in order to provide better speed and increase customer conversion by employing a computational module. The system segments the specific customer cohorts based on the user attributes, product attributes and transaction type which leads to a differentiated service experience to the specific customer cohorts. The system (100) for predictive interventions in capacity allocations, comprising a predictive unit and a reactive unit wherein the predictive unit comprises of a forecasting module (101) that comprises a multiple cohorts, a historical demand data, an inventory data, a macroeconomic data, a weather data and a set of promotions; an optimization module (102) comprises a cohort priority, a spillage budget in order to give better protection and a set of cut-offs, fed from the forecasting module (101); a capacity allocation module (103) connected to the optimization module (102); and the reactive unit comprises a guardrail unit (104) that is connected with the capacity allocation module (103) and runs on top of the forecasting module and the optimization module (102). The forecasting module (101) predicts the demand for the capacity and resources for the multiple cohorts during a predefined event. The optimization module (102) finds an optimal way to allocate the capacity and resources of the cohort priority based on the forecast distribution in the forecasting module (101) across the multiple cohorts and the set of cut-offs in the predefined duration to minimize spillage, thereby mitigating the risk of spillage protection and under-utilization. The guardrail unit (104) sets rules to guide the system (100) in optimizing the utilization of resources. The capacity allocation module (103) manages the dynamic re-allocation of the capacity and resources on approval of the multiple cohorts.

DETAILED DESCRIPTION OF INVENTION

[0023] The following detailed description and embodiments set forth herein below are merely exemplary out of the wide variety and arrangement of instructions which can be employed with the present invention. The present invention may be embodied in other specific forms without departing from the spirit or essential characteristics thereof. All the features disclosed in this specification may be replaced by similar other or alternative features performing similar or same or equivalent purposes. Thus, unless expressly stated otherwise, they all are within the scope of the present invention.

[0024] Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope of the invention. In addition, descriptions of well-known functions and constructions are omitted for clarity and conciseness.

[0025] It should be emphasized that the term “comprises/comprising” when used in this specification is taken to specify the presence of stated features, steps or components but does not preclude the presence or addition of one or more other features, steps, components or groups thereof.

[0026] The present invention relates to a system for predictive interventions in capacity allocations to reserve capacity for the selected cohort to avoid spillage, thereby protecting priority cohorts from speed degradation.

[0027] In an embodiment, the present invention provides a system (100) for predictive interventions in capacity allocations, comprising a predictive unit and a reactive unit wherein the predictive unit comprises of a forecasting module (101) that comprises a multiple cohorts, a historical demand data, an inventory data, a macroeconomic data, a weather data and a set of promotions; an optimization module (102) comprises a cohort priority, a spillage budget in order to give better protection and a set of cut-offs, fed from the forecasting module (101); a capacity allocation module (103) connected to the optimization module (102); and the reactive unit comprises a guardrail unit (104) that is connected with the capacity allocation module (103) and runs on top of the forecasting module (101) and the optimization module (102). The system is flexible to be utilized during any part of the year. The forecasting module (101) predicts the demand for the capacity and resources for the multiple cohorts during a predefined event. The optimization module (102) finds an optimal way to allocate the capacity and resources of the cohort priority based on the forecast distribution in the forecasting module (101) across the multiple cohorts and the set of cut-offs in the predefined duration to minimize spillage, thereby mitigating the risk of spillage protection and under-utilization. The optimization module (102) employs a stochastic modeling approach. The guardrail unit (104) sets rules to guide the system (100) in optimizing the utilization of resources. The capacity allocation module (103) manages the dynamic re-allocation of the capacity and resources on approval of the multiple cohorts.

[0028] In another embodiment, the set of promotions are employed in order to provide discounts. The system further interacts with a promise engine to define the capacity at source by employing a multi- cohort scenario in accordance with the demand of the customers to protect a plurality of specific cohorts from speed degradation.

[0029] In another embodiment, the present invention provides a system (100) to create capacity reservations in supply chain in order to prioritize orders from protected cohorts to provide better speed.

[0030] (a) Spillage and Capacity Consumption:
(I) Single Cohort Scenario: When a customer places an order on a portal, a Promise engine is invoked to determine the (expected) delivery SLA for that order. To determine the SLA, the promise engine first identifies the closest source with available inventory. The promise engine then finds the first available wave at the source to which the order is assigned to (a source operates in waves where each wave is associated with a processing capacity and wave end time). This enables the Promise engine to determine the time at which the order is dispatched from the source, which ultimately helps compute the delivery SLA to be shown to the customer. When the source capacity is lower than demand, orders spill to a future wave resulting in a degraded SLA to the customer. This problem is significant during Events (e.g., Republic Day Sale) and BBD (Big Billion Day Sale).

[0031] Figure 1 illustrates an exemplary flow chart to demonstrate spillage view in single cohort scenario. The cut-off, CO-1 is at time point t3 with 600 units capacity while the total demand till t3 is 300+600 = 900 units. The numbers above the timeline show the demand for each time period. Spillage is the number of demand units that overflow to the next cut-off when capacity is insufficient. This results in 300 orders spilling to the next cut-off. Such a cascading effect ends up degrading the promised speed. Similarly, the cut-off CO-2 is at the time point t6 with 900 units capacity while the total demand till t6 is 300+500+300+500 = 1600 units, which results in the spillage of 700 (which included spillage from CO-1).

[0032] (II) Multi-Cohort Scenario: To avoid spillage of orders in the selected cohort, Business team creates multiple capacity pools and each pool is only used by specified orders. Figure 2 illustrates an exemplary flowchart of cut-offs and the case of two capacity pools (i.e. Non-Protected and Protected). Spillage for each pool is computed based on their respective demand & assigned capacities to demonstrate spillage view in multiple cohorts as shown in Figure, wherein 100 orders are approved for protected cohort in each time point. Capacity of 200 units allocated for cut-off point-1 (at time point t3) and 300 units in cut-off point-2 at time point t6 are enough to avoid spillage of the protected cohort. Promise while approving orders of protected cohorts consume capacity reserved exclusively to this cohort (assuming disjoint capacity consumption scenario). As these orders are processed in the same cut-off, they are connected to ideal LPHTs (earliest possible connections) enabling sources to dispatch orders on the same day leading to significant speed goodness. Business team has flexibility to create multiple cohorts based on desired speed offering to the defined cohorts. The multiple cohorts are segmented on the basis of user attributes, product attributes, supply chain network attributes and transaction type.
Protection % measured at a cut-off level or a day-level. At a day-level, it is defined as the % of orders that is processed and dispatched from the source on the same day. Protection directly impacts the speed that we offer to the customers. Total protection is defined as the % of orders in the entire event that had a same-day dispatch from the source.
Protection % = 100* (Count of orders with same-day dispatch/ Count of orders during the event) (1)
Spillage % corresponds to the % of orders that spilled beyond the recovery period, i.e., the % of orders that could not be dispatched from the FC during the “Event +Recovery" period.
Spillage % = 100* (Count of orders that spilled beyond recovery period/ Count of Orders during Event+ Recovery period) (2)

[0033] (b) Under-utilization: Capacities for all cohorts have to be defined for each cut-off beforehand, it may be possible that realized demands are less than the reserved cohort capacities. Figure 3 illustrates an exemplary flowchart to demonstrate under-utilization view for non-protected and protected view. For the first cut-off at time-point 𝑡3, the non-protected cohort is experiencing a spillage of 300 units. At the same time, the demand realized for the protected cohort (50+50 units) is lower than the reserved capacity of 200 which results in 100 units of capacity under-utilization. This capacity if provided to the non-protected cohort could have reduced its spillage. The 100 units of capacity is thus considered as under-utilization induced by the capacity segmentation. On the other hand, for the cut-off at time point 𝑡6, both the cohorts have their demand lower than the available capacity. In this case, there is an overall under attainment of the demand which is not caused by capacity segmentation. This is not considered as under-utilization in this work.
Under-utilization % for entire event is defined as a percentage of total capacity
Under-util % = 100* (Total under-util of the event/ Total capacity) (3)

[0034] (c) Cut-offs & capacities: List of planned cut-offs timings for the event with their allocated capacities.

[0035] (d) Under-utilization budget: Business team provides an upper bound on the system induced under-utilization. This is generally defined as a percentage of total capacity. As an example, the under-utilization budget is in the range of 0.5% to 1.5% of the total event capacity.

[0036] (e) Spillage budget: Allowed spillage after the end of recovery period. This is BAU spillage plus allowed budget (linked to under-utilization). Ex: If BAU spillage is 10000 and the under-utilization allowed is 0.5% of the total capacity (say, 100,000), then the allowed spillage is 10000+0.005*100000 = 10,500 units.

[0037] (f) Cohort Definitions: A cohort is typically defined using one or more of the below pivots/filters. The system is flexible to be utilized during any part of the year based on cohort definitions:
– Customer type (e.g., VIP, Plus)
– Business unit (e.g., BGM, Electronics, Lifestyle)
– Fulfilment tier (e.g., Express, Economy, Regular)
– LZN flag (e.g., local, zonal, national)
– City tier (e.g., Metro, Tier1, Tier2)
– Dangerous flag – Average Selling Price bucket (e.g., 500-1000, 1000-3000, >3000)
– Carrier type (e.g., Ekart, 3PL)
– Volumenous flag
-Supply chain network attributes example based on customer pin codes

[0038] (g) Cohort Priority: Defines the priority order for speed protection among the defined cohorts.

[0039] (h) Cohort X Cut-off capacities: Provide optimal allocation of cut-off capacity to all cohorts.

[0040] Figure 4 illustrates the system for predictive unit and reactive unit in capacity allocations.

[0041] Forecasting Module (101): The Forecasting module (101) forecasts demand for different combinations of cut-offs and the multiple cohorts by analyzing the historical demand data. The Forecasting module (101) employs a probabilistic modeling approach or a point demand forecast model. Forecasting module (101) is to obtain forecast distribution for all cohorts X cut-offs of the event. Due to promotions/offers, variability of demand at a cut-off level is very high and relying on a point estimate leads to loss of protection or under-utilization in the cohorts. To handle such situation, forecasting module (101) provides 3 point distribution. Each time these optimization module (102) triggered, the demand forecasts till the end of the event are required as inputs. A probabilistic demand forecast model which outputs discrete probability distributions of demands till the immediate n-cutoffs at a cutoff level. A point demand forecast model which predicts demands for each cutoff for the duration of the event after the immediate n-cutoffs.

[0042] Optimization Module (102): Optimization module (102) provides optimal capacity allocation across cohorts for all cut-offs by taking forecast distribution, spillage constraints & cut-off capacities in order to maximize protection i.e. minimize spillage for selected (high priority) cohorts for the entire event. Therefore, a stochastic modeling approach is employed by creating possible scenarios from forecast distribution.

[0043] Guardrail Module: The guardrail module works on the capacity allocation module (103). The essence of the guardrail module is to define a framework in the interest of the platform objectives while allowing the system to do its job within that framework, so as to bring in certainty and predictability in the outcome of the system.
• Capacity under-utilisation within allowed budget
• Limiting spillage of non-protected pivots and in turn providing good SLAs
• Protect overall speed coverage numbers as D0, D1, D2, etc.
Based on above objectives, the guardrail module runs on the top of forecasting & optimization modules to minimize the system induced underutilization. Based on utilization of cohorts, move capacity from protected cohorts to non-protected cohorts. It is a rule-based system as of now and main metric it considers is under-utilization of a cohort in a cut-off which is defined as below:
Underutilization in a cut-off = 100* (1 − Demand in the Cut-off/ Capacity in the Cut-off )

[0044] Capacity allocation module (103): Capacity consumption during approval/promise is based on the utilization strategy. Some possible strategies are detailed below:
• FCFS (First come First Serve): Create a single capacity pool which gets utilized in a FCFS fashion (i.e., no priority to any cohort). This strategy does not offer any speed protection, nor does it result in capacity underutilization.
• Disjoint Strategy: Create disjoint capacity pools per cohort. Each cohort is strictly restricted to use capacity from its own pool. This strategy enables speed protection to high priority cohorts and lead to capacity underutilization as well. It can also lead to spillage/higher SLAs for Non-protected cohort.
• Nested Strategy: Create nested capacity pools. For example, a dedicated pool that can only be used by the protected cohort, and a common pool that can be used by any. Protected cohort consume their capacity first and then use the common pool. Non-protected uses capacity only from common pool.

[0045] In an exemplary embodiment, Figure 5 illustrates a graphical representation of Demand, Capacity and Spillage for a major event with event days & recovery and at the end of recovery days, spillage is expected to be at BAU level number. The major events (tier-1, tier-2 & BBD) range from 5 to 9 days. Typically, an event starts with an early access (before the day of the event start) to VIP/Plus customers & remains open to all customers from Day-1 of the event. Most of the promotions are provided on first 2-3 days of the sale and hence, demand spikes during these days. As total capacity of first few days will generally be lesser than realized demand, spillage spikes in first few days and continues till end of event. Typically, business team plans recovery days post event to control the spillage caused during event to BAU spillage. It was a 9-day event with 2 days planned for recovery. Note that the demand is high on the first 2-3 days of the event and on the last day of the event (day 9). This leads to significant spillages that continues till the end of the event and reaches the BAU levels only after the end of the recovery period.

[0046] In another exemplary embodiment, Figure 6 illustrates a block diagram of simulation engine with inputs & outputs and interaction with three modules Forecasting, Optimization and Guardrail. The system’s performance depends on integration of three modules and other external factors such as actual order profile/pattern, consumer behavior with promotions, inventory placement/availability in sources, external competition, etc. External factors for every event is different and modeling them is very complex. Even with our internal/controllable factors, it gets challenging to validate any proposed solution due to unavailability of a closed-form solution primarily due to unknown demand pattern. Hence, we developed a simulation engine to validate the proposed solution & tested on multiple past events and obtain performance metrics. For testing, we considered top tier-1 events & BBD’22 and for every test set, we compared metrics of PICASA run with FCFS strategy (a baseline without PICASA). We ran simulation on top 4 FCs that contribute to 25% of FBF volume. An overview of Simulation engine implementation is provided below:
• Get historic order profile (order level data) with following attributes
– OrderID
– Order approval time (in datetime)
– Order Cohort type (Eg. VIP, High-ASP, etc.)
• Get Source capacity of considered FC that provides data of Cut-off’s of the event & respective available capacity.
• We ran simulation in multiple settings of capacity consumption.
They are
– FCFS (No priority in capacity consumption)
– Strategy-3 (Nested strategy: protected cohort will consume their capacity first, and then use the common pool)
• Obtain forecasts of all cut-offs of the event before start of the event and run optimization model to allocate capacities
• Iterate over each order (in the order of their approval time) and identify earliest cut-off with available capacity based on selected strategy & consume its capacity
• Before P2D time of each cut-off, Both forecasting & optimization model are re-run to re-allocate capacities
• Every hour (at 15th minute), Guardrail runs to move capacity from protected cohorts to common pool to reduce underutilization.

[0047] Figure 7 (a) illustrates a graphical representation of comparison of Spillage between FCFS and model results for 4 major FCs/ Warehouses for a major event. Event is of 9 days with a recovery of 2 days. This indicates by the end of the event plus 2 days, spillage of FCs should be near to BAU spillages. Order profiles & Source capacity data were created and ran simulation runs on these 4 FCs. Protection & Spillage are compared for FCFS & Nested strategy (model).
[0048] Figure 7 (b) illustrates a graphical representation of comparison of protection provided to VIP cohort between FCFS and Model results for 4 major FCs/Warehouses. In comparison to FCFS, protection provided to VIP cohort has increased from 15% to around 75% in FC1, FC2, FC3 & FC4. In FC3, it is improved 55% due to higher spillages & high mapes in forecasting.
[0049] Figure 7 (c) illustrates a graphical representation of comparison of protection provided to High-ASP cohort between FCFS and Model results for 4 major FCs/Warehouses. In comparison to FCFS, protection provided to High-ASP has increased from <10% to >50% in all FCs.
[0050] Spillage in all FCs are within thresholds set during runs (and these were set based on allowed spillages in respective FCs).
[0051] While this invention has been described in connection with what is presently considered to be the most practical and preferred embodiment, it is to be understood that the invention is not limited to the disclosed embodiments, but, on the contrary, is intended to cover various modifications and equivalent arrangements included within the scope of the appended claims.
, Claims:1. A system (100) for dynamic capacity orchestration in a segmented supply chain to maximize customer experience, comprising:
• a predictive unit, comprising:
o a forecasting module (101) comprises multiple cohorts, a historical demand data, an inventory data, a macroeconomic data, a weather data and a set of promotions;
o an optimization module (102) comprises a cohort priority, a spillage budget and a set of cut-offs, fed from the forecasting module (101);
o a capacity allocation module (103) connected to the optimization module (102); and
• a reactive unit, comprising
o a guardrail unit (104) connected with the capacity allocation module (103) and runs on top of the forecasting module (101) and the optimization module;
wherein,

• the forecasting module (101) predicts the demand for the capacity and resources for the multiple cohorts during a predefined event;
• the optimization module (102) finds an optimal way to allocate the capacity and resources of the cohort priority based on the forecast distribution in the forecasting module (101) across the multiple cohorts and the set of cut-offs in the predefined duration to minimize spillage, thereby mitigating the risk of spillage protection and under-utilization;
• the guardrail unit (104) sets rules to guide the system (100) in optimizing the utilization of resources; and
• the capacity allocation module (103) manages the dynamic re-allocation of the capacity and resources on approval of the multiple cohorts.

2. The system (100) for dynamic capacity orchestration in a segmented supply chain to maximize customer experience as claimed in claim 1, wherein the set of promotions are employed in order to provide discounts.

3. The system (100) for dynamic capacity orchestration in a segmented supply chain to maximize customer experience as claimed in claim 1, wherein the multiple cohorts are segmented on the basis of user attributes, product attributes, supply chain network attributes and transaction type.

4. The system (100) for dynamic capacity orchestration in a segmented supply chain to maximize customer experience as claimed in claim 1, wherein the forecasting module (101) employs a probabilistic modeling approach or a point demand forecast model.

5. The system (100) for dynamic capacity orchestration in a segmented supply chain to maximize customer experience as claimed in claim 1, wherein the optimization module (102) employs a stochastic modeling approach.

6. The system (100) for dynamic capacity orchestration in a segmented supply chain to maximize customer experience as claimed in claim 1, wherein the forecasting module (101) forecasts demand for different combinations of cut-offs and the multiple cohorts by analyzing the historical demand data.

7. The system (100) for dynamic capacity orchestration in a segmented supply chain to maximize customer experience as claimed in claim 1, wherein the system interacts with a promise engine to define the capacity at source by employing a multi- cohort scenario in accordance with the demand of the customers to protect a plurality of specific cohorts from speed degradation.

8. The system (100) for dynamic capacity orchestration in a segmented supply chain to maximize customer experience as claimed in claim 1, wherein the system is flexible to be utilized during any part of the year based on cohort definitions.

9. The system (100) for dynamic capacity orchestration in a segmented supply chain to maximize customer experience as claimed in claim 1, wherein the cohort priority defines the priority order for speed protection among the defined cohorts.