Area based delivery order notification and management

An apparatus, comprising:

• a communication interface communicatively connected to one or more nodes of a system of nodes;
• and
• a processor configured to:
• receive, at the apparatus, an initial order to be physically fulfilled by one or more nodes of the system of nodes, wherein the initial order specifies a location and quantity parameters of fulfilment;
• generate a fulfilment criterion for the initial order using a first machine learning predictive model, wherein the first machine learning predictive model is trained upon data from prior fulfilment operations to generate predictions on future fulfilment criterion, wherein the fulfilment criterion is based on the quantity parameters of fulfilment;
• generate a local fulfilment constraint for the initial order using a second machine learning predictive model, wherein the second machine learning predictive model is trained to generate baseline predictions of an effect produced by the local fulfilment constraint;
• generate, based on at least one of the initial order, the fulfilment criterion for the initial order, and the local fulfilment constraint for the initial order, one or more conditional optimization opportunities, wherein each of the one or more conditional optimization opportunities comprises an action, which if taken, improves one or more performance metrics of the system of nodes;
• generate a conditional input for optimization, wherein the conditional input for optimization defines an additional task to be taken by a node of the system;
• send the conditional input for optimization to a plurality of candidate nodes of the system via the communication interface, the plurality of candidate nodes identified based on satisfying the local fulfillment constraint generated for the initial order;
• determine a dynamic optimization window based on the fulfilment criterion and the local fulfilment constraint, the dynamic optimization window defining a period of time during which responses from the plurality of candidate nodes are receivable by the apparatus;
• receive, within the dynamic optimization window, a response from at least one candidate node of the plurality of candidate nodes, wherein the response satisfies a condition of the conditional input for optimization;
• generate based at least in part on the response from the at least one candidate node, an optimized set of control inputs, wherein the optimized set of control inputs define actions to be taken by nodes to physically fulfil the initial order and perform the additional task to realize at least one of the one or more conditional optimization opportunities;
• and
• send the optimized set of control inputs to the nodes for physically fulfilling the initial order and realizing the at least one conditional optimization opportunities.

A non-transitory, computer-readable medium containing code that, when executed by a processor, causes an apparatus to:

• receive, at the apparatus, an initial order to be physically fulfilled by one or more nodes of a system of nodes communicatively connected to the apparatus, wherein the initial order specifies a location and quantity parameters of fulfilment;
• generate a fulfilment criterion for the initial order using a first machine learning predictive model, wherein the first machine learning predictive model is trained upon data from prior fulfilment operations to generate predictions on future fulfilment criterion, wherein the fulfilment criterion is based on the quantity parameters of fulfilment;
• generate a local fulfilment constraint for the initial order using a second machine learning predictive model, wherein the second machine learning predictive model is trained to generate baseline predictions of an effect produced by the local fulfilment constraint;
• generate, based on at least one of the initial order, the fulfilment criterion for the initial order, and the local fulfilment constraint for the initial order, one or more conditional optimization opportunities, wherein each of the one or more conditional optimization opportunities comprises an action, which if taken, improves one or more performance metrics of the system of nodes;
• generate a conditional input for optimization, wherein the conditional input for optimization defines an additional task to be taken by a node of the system;
• send the conditional input for optimization to a plurality of candidate nodes of the system of nodes, the plurality of candidate nodes identified based on satisfying the local fulfillment constraint generated for the initial order;
• determine a dynamic optimization window based on the fulfilment criterion and the local fulfilment constraint, the dynamic optimization window defining a period of time during which responses from the plurality of candidate nodes are receivable by the apparatus;
• receive, within the dynamic optimization window, a response from at least one candidate node of the plurality of candidate nodes, wherein the response satisfies a condition of the conditional input for optimization;
• generate, based at least in part on the response from the at least one candidate node, an optimized set of control inputs, wherein the optimized set of control inputs define actions to be taken by nodes to physically fulfil the initial order and perform the additional task to realize at least one of the one or more conditional optimization opportunities;
• and
• send the optimized set of control inputs to the nodes for physically fulfilling the initial order and realizing the at least one conditional optimization opportunities.