Abstract: Techniques for training a prediction model are disclosed. An example method includes processing historical event data comprising a plurality of computer-readable events for a resource to determine previous parameters for the resource. The method also includes generating training data for training the prediction model without using a forecast for future utilization of the resource. The training data comprises a set of proxies generated from previous parameters for the resource. Each proxy is associated with a remaining capacity of the resource and a remaining time to expiration of the resource. The method also includes training the prediction model to generate a mapping from the remaining capacity of the resource and the remaining time to expiration of the resource to the proxy. The method also includes receiving a request that describes a potential future event pertaining to the resource and generating a prediction for the potential future event using the prediction model.

Abstract: Systems and methods for network optimization in a distributed big data environment are provided. According to an aspect of the invention, a processor performs an optimization method by dividing a data set into a plurality of partitions. For each of the partitions, the processor generates a mathematical representation of a model by associating input data with elements of the model, wherein the mathematical representation includes an objective and at least one constraint. The processor forms a master objective by combining the objectives for the partitions, and forms a set of master constraints by combining the constraints for the partitions. The processor then generates an optimized solution based on the master objective and the master constraints.

Abstract: Systems and methods for network optimization in a distributed big data environment are provided. According to an aspect of the invention, a processor performs an optimization method by dividing a data set into a plurality of partitions. For each of the partitions, the processor generates a mathematical representation of a model by associating input data with elements of the model, wherein the mathematical representation includes an objective and at least one constraint. The processor forms a master objective by combining the objectives for the partitions, and forms a set of master constraints by combining the constraints for the partitions. The processor then generates an optimized solution based on the master objective and the master constraints.

Abstract: Techniques are disclosed for adjusting inventory allocations assigned to system elements published to a reservation and booking system. In particular, inventory allocations assigned to a system element may be adjusted based on deviations from a target utilization level assigned to that element. For example, in the context of an airline pricing and reservation system, a threshold revenue management (RM) system may adjust fare class level (FCL) allocations for a published flight. As disclosed, the threshold RM system may make across-run adjustments when a flight is published in a booking system and across-DCP adjustments after the flight is published. Doing so may both improve revenue as well as help ensure a flight realizes an assigned load factor target at departure.

Abstract: Techniques are disclosed for propagating a system level utilization goal to individual system elements. For example, a utilization goal such as a load factor goal for a group of airline flights in a given market segment and time period may be propagated to each flight in the group. When propagating a load factor goal to a group of airline flights, a goal value, historical load factors, and a capacity of each flight in the group may be used to determine a load factor target for each flight in the flight group. The propagated load factor targets are expected to be realizable by each flight in the flight group and, in the aggregate, satisfy the system level goal.

Abstract: Systems and methods for optimizing marketing strategies. Various embodiments implement methods which can include generating a plurality of candidate solutions which satisfy pricing rationality constraints. The candidate solutions can be generated when processing resources are available for performing a Monte Carlo algorithm. The candidate solutions can be stored and a master and trade off metric can be selected. Values for these selected metrics can be evaluated (at the candidate solutions) and then input into the approximate efficient frontier algorithm. The algorithm can output an approximate efficient frontier. Users can select any of the efficient solutions on the frontier to obtain associated pricing recommendations by mousing over the efficient frontier. Various metrics associated with the efficient frontier may be updated at about the same time as the master and trade off metrics are selected.