System and method for setting inventory thresholds for offering and fulfillment across retail supply networks

A computer system for determining whether to offer an item for sale on line or whether to source an item from one or more fulfillment nodes, comprising:

• a processor;
• and
• a memory having program instructions, which when executed by the processor cause a probabilistic cancellation module to perform the steps of:
• receiving by a computer processor a electronic record of a request to offer an item for sale or an order from a customer for the item, the computer processor having program instructions, which when executed by the processor, performs the steps of:
• retrieving historical data and current data, the historical data and current data comprising inventory data at order creation time for the item, inventory data at node source assignment time for the item, one or more item features of the item, the one or more item features comprising at least one of cost of the item and weight of the item, and one or more node features of one or more nodes from a plurality of nodes, the one or more node features comprising at least one of node capacity utilization and a cancelation rate;
• deriving and training a probabilistic model of cancellation on the historical data using machine learning, the probabilistic model of cancellation being a logistic regression model represented by a sigmoidal function;
• training the probabilistic model of cancellation on updated historical data using machine learning;
• deriving data from the probabilistic model of cancellation, the derived data including inventory derived values, the inventory derived values comprising at least one of predicted days and weeks of supply;
• inputting the current data into the probabilistic model of cancellation;
• automatically determining, by the probabilistic model of cancellation, using the derived data, the retrieved historical data and current data, a probability of cancellation of the item;
• dynamically generating, by the probabilistic model of cancellation, an inventory threshold rule for offering the item for sale or an inventory threshold rule for sourcing the item, when the probability of cancellation of the item is lower than a predetermined cancelation rate of the item;
• in a case when the probability of cancellation of the item is lower than a predetermined cancelation rate of the item, generating, by the computer processor, an electronic message regarding offering the item for sale on an on-line website when the current inventory for the item is above the inventory threshold, or sourcing the item from one of the plurality of nodes, when the current inventory for the item is above the inventory threshold;
• and
• transmitting the message regarding offering the item for sale to the on-line website or transmitting the message regarding sourcing the item to an order fulfillment engine.

A non-transitory article of manufacture tangibly embodying computer readable instructions, which when implemented, cause a computer to perform the steps of a method for determining whether to offer an item for sale on line or whether to source an item from one or more fulfillment nodes, comprising:

• receiving by a computer processor a electronic record of a request to offer an item for sale or an order from a customer for the item, the computer processor having program instructions, which when executed by the processor, performs the steps of:
• retrieving historical data and current data, the historical data and current data comprising inventory data at order creation time for the item, inventory data at node source assignment time for the item, one or more item features of the item, the one or more item features comprising at least one of cost of the item and weight of the item, and one or more node features of one or more nodes from a plurality of nodes, the one or more node features comprising at least one of node capacity utilization and a cancelation rate;
• deriving and training a probabilistic model of cancellation on the historical data using machine learning, the probabilistic model of cancellation being a logistic regression model represented by a sigmoidal function;
• training the probabilistic model of cancellation on updated historical data using machine learning;
• deriving data from the probabilistic model of cancellation, the derived data including inventory derived values, the inventory derived values comprising at least one of predicted days and weeks of supply;
• inputting the current data into the probabilistic model of cancellation;
• automatically determining, by the probabilistic model of cancellation, using the derived data, the retrieved historical data and current data, into a probability of cancellation of the item;
• dynamically generating, by the probabilistic model of cancellation, an inventory threshold rule for offering the item for sale or an inventory threshold rule for sourcing the item, when the probability of cancellation of the item is lower than a predetermined cancelation rate of the item;
• in a case when the probability of cancellation of the item is lower than a predetermined cancelation rate of the item, generating, by the computer processor, an electronic message regarding offering the item for sale on an on-line website when the current inventory for the item is above the inventory threshold, or sourcing the item from one of the plurality of nodes, when the current inventory for the item is above the inventory threshold;
• and
• transmitting the message regarding offering the item for sale to the on-line website or transmitting the message regarding sourcing the item to an order fulfillment engine.