Utilizing a learning engine in predicting physical resource utilization

Abstract

Methods and systems are disclosed for allocating physical resources based on predicted utilization demand utilizing a learning engine. Historical resource reservation data for a physical resource is accessed, wherein the physical resource is shared amongst physical resource users in a time displaced manner. Historical event data, comprising event types and event dates, is accessed. Correlations between historical reservation data and the historical event data are identified and are used by a learning engine to predict future high demand dates. A notification is generated regarding a predicted future high demand date, including a control via which a request may be initiated for the physical resource for the first predicted future high demand date. The notification may be transmitted to physical resource user(s), and in response to a detection of an activation of the request initiation control, a corresponding allocation for the physical resource is registered and the physical resource may be accordingly utilized.

What is claimed is: 
     
         1 . A system configured to enable allocation of physical resources, comprising:
 a network interface;   at least one processing device operable to:
 access historical utilization data, comprising occupancy data, for a first physical resource, the historical utilization data comprising dates at which the first physical resource was utilized by a given user, wherein the first physical resource is shared amongst a plurality of physical resource users in a time displaced manner; 
 access historical event data, the historical event date comprising event types and event dates; 
 use a neural network comprising an error function to identify correlations between historical utilization data, comprising occupancy data, and the historical event data, comprising event types and event dates; 
 based at least in part on the identified correlations between historical utilization data, comprising occupancy data, and the historical event data, comprising event types and event dates, predict future high utilization dates for the first physical resource; 
 generate a notification regarding at least a first predicted future high utilization date, the notification including a control via which a reservation may be initiated for the first physical resource for the first predicted future high utilization date; 
 transmit the notification regarding the first predicted future high utilization date to at least one physical resource user in the plurality of physical resources users; 
 at least partly in response to a detection of an activation of the control via which a reservation may be initiated for the first physical resource for the first predicted future high utilization date, registering a corresponding reservation for the first physical resource for the first predicted future high utilization date. 
   
     
     
         2 . The system as defined in  claim 1 , wherein the neural network comprises:
 one or more convolution layers and/or a multi-layer long short-term memory component.   
     
     
         3 . The system as defined in  claim 1 , wherein the historical utilization data comprises waiting list data. 
     
     
         4 . The system as defined in  claim 1 , wherein the historical utilization data comprises requests for reservations for the first physical resource. 
     
     
         5 . The system as defined in  claim 1 , wherein the historical event data comprises events associated with a geographical region where the physical resource is located. 
     
     
         6 . The system as defined in  claim 1 , the operations further comprising:
 identify which physical resource user in the plurality of physical resources meets a first eligibility criterion, the first eligibility criterion relating to historical reservations for the first physical resource;   wherein the notification regarding the first predicted future high utilization date is initially transmitted to a first physical resource user in the plurality of physical resources that satisfies the first eligibility criterion; and   at least partly in response to failing to detect a reservation by the first physical resource user within a first threshold period of time, transmit the notification regarding the first predicted future high utilization date to one or more other physical resource users.   
     
     
         7 . The system as defined in  claim 1 , wherein the notification regarding the first predicted future high utilization date is initially transmitted to a first physical resource user in the plurality of physical resources users, wherein the system is configured to, at least partly in response to failing to detect a reservation by the first physical resource user within a first threshold period of time, transmit the notification regarding the first predicted future high utilization date to one or more other physical resource users. 
     
     
         8 . The system as defined in  claim 1 , wherein the notification regarding the first predicted future high utilization date is transmitted to each of the plurality of physical resources users. 
     
     
         9 . The system as defined in  claim 1 , wherein the notification regarding the first predicted future high utilization date identifies an event correlated with the first predicted future high utilization date. 
     
     
         10 . The system as defined in  claim 1 , wherein the first physical resource comprises an inhabitable structure. 
     
     
         11 . A computer-implemented method, the method comprising:
 using a computer system comprising one or more computer devices, accessing historical demand data comprising reservation data for a first physical resource, the historical demand data comprising dates at which the first physical resource was reserved by a given user, wherein the first physical resource is shared amongst a plurality of physical resource users in a time displaced manner;   accessing, using the computer system, historical event data, the historical event date comprising event types and event dates;   identifying, using the computer system, correlations between historical demand data, comprising reservation data, and the historical event data, comprising event types and event dates;   based at least in part on the identified correlations between historical demand data, comprising reservation data, and the historical event data, comprising event types and event dates, predicting future high demand dates for the first physical resource;   generating a notification regarding at least a first predicted future high demand date, the notification including a control via which a reservation may be initiated for the first physical resource for the first predicted future high demand date;   transmitting the notification regarding the first predicted future high demand date to at least one physical resource user in the plurality of physical resources users;   at least partly in response to a detection of an activation of the control via which a reservation may be initiated for the first physical resource for the first predicted future high demand date, registering a corresponding reservation for the first physical resource for the first predicted future high demand date.   
     
     
         12 . The method as defined in  claim 11 , wherein identifying correlations between historical demand data, comprising reservation data, and the historical event data, comprising event types and event dates, is performed using a neural network comprising:
 one or more convolution layers and/or a multi-layer long short-term memory component.   
     
     
         13 . The method as defined in  claim 11 , wherein the historical demand data comprises waiting list data. 
     
     
         14 . The method as defined in  claim 11 , wherein the historical demand data comprises requests for reservations for the first physical resource. 
     
     
         15 . The method as defined in  claim 11 , wherein the historical event data comprises events associated with a geographical region where the physical resource is located. 
     
     
         16 . The method as defined in  claim 11 , the method further comprising:
 identifying which physical resource user in the plurality of physical resources meets a first eligibility criterion, the first eligibility criterion relating to historical reservations for the first physical resource;   wherein the notification regarding the first predicted future high demand date is initially transmitted to a first physical resource user in the plurality of physical resources that meets the first eligibility criterion; and   at least partly in response to failing to detect a reservation by the first physical resource user within a first threshold period of time, transmitting a subsequent notification regarding the first predicted future high demand date to one or more other physical resource users.   
     
     
         17 . The method as defined in  claim 11 , wherein the notification regarding the first predicted future high demand date is initially transmitted to a first physical resource user in the plurality of physical resources users, the method further comprising: at least partly in response to failing to detect a reservation by the first physical resource user within a first threshold period of time, transmitting the notification regarding the first predicted future high demand date to one or more other physical resource users. 
     
     
         18 . The method as defined in  claim 11 , wherein the notification regarding the first predicted future high demand date is transmitted to each of the plurality of physical resources users. 
     
     
         19 . The method as defined in  claim 11 , wherein the notification regarding the first predicted future high demand date identifies an event correlated with the first predicted future high demand date. 
     
     
         20 . Non-transitory computer readable memory having program instructions stored thereon that when executed by a computer system device cause the computer system to perform operations comprising:
 access historical demand data, comprising reservation data, for a first physical resource, the historical demand data comprising dates at which the first physical resource was reserved by a given user, wherein the first physical resource is shared amongst a plurality of physical resource users in a time displaced manner;   access historical event data, the historical event date comprising event types and event dates;   identify correlations between historical demand data, comprising reservation data, and the historical event data, comprising event types and event dates;   predict future high demand dates for the first physical resource based at least in part on the identified correlations between historical demand data, comprising reservation data, and the historical event data, comprising event types and event dates;   generate a notification regarding a first predicted future high demand date;   transmit the notification regarding the first predicted future high demand date to at least one physical resource user in the plurality of physical resources users;   in response to a reservation request, register a corresponding reservation for the first physical resource for the first predicted future high demand date.   
     
     
         21 . The non-transitory computer readable memory as defined in  claim 20 , wherein identifying correlations between historical demand data, comprising reservation data, and the historical event data, comprising event types and event dates, is performed using a neural network comprising:
 one or more convolution layers and/or a multi-layer long short-term memory component.   
     
     
         22 . The non-transitory computer readable memory as defined in  claim 20 , wherein the historical demand data comprises waiting list data. 
     
     
         23 . The non-transitory computer readable memory as defined in  claim 20 , wherein the historical demand data comprises requests for reservations for the first physical resource. 
     
     
         24 . The non-transitory computer readable memory as defined in  claim 20 , wherein the historical event data comprises events associated with a geographical region where the physical resource is located. 
     
     
         25 . The non-transitory computer readable memory as defined in  claim 20 , the operations further comprising:
 identify which physical resource user in the plurality of physical resources meets a first eligibility criterion, the first eligibility criterion relating to historical reservations for the first physical resource;   wherein the notification regarding the first predicted future high demand date is initially transmitted to a first physical resource user in the plurality of physical resources that meets the first eligibility criterion; and   at least partly in response to failing to detect a reservation by the first physical resource user within a first threshold period of time, transmit the notification regarding the first predicted future high demand date to one or more other physical resource users.   
     
     
         26 . The non-transitory computer readable memory as defined in  claim 20 , wherein the notification regarding the first predicted future high demand date is initially transmitted to a first physical resource user in the plurality of physical resources users, the method further comprising: at least partly in response to failing to detect a reservation by the first physical resource user within a first threshold period of time, transmitting the notification regarding the first predicted future high demand date to one or more other physical resource users. 
     
     
         27 . The non-transitory computer readable memory as defined in  claim 20 , wherein the notification regarding the first predicted future high demand date is transmitted to each of the plurality of physical resources users. 
     
     
         28 . The non-transitory computer readable memory as defined in  claim 20 , wherein the notification regarding the first predicted future high demand date identifies an event correlated with the first predicted future high demand date.