Browse Prior Art Database

System for Coordination of Recommendations to Reduce Resource Contention

IP.com Disclosure Number: IPCOM000237218D
Publication Date: 2014-Jun-09
Document File: 2 page(s) / 21K

Publishing Venue

The IP.com Prior Art Database

Abstract

A system and method for coordination of recommendations by a service to reduce resource contention is disclosed.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 51% of the total text.

Page 01 of 2

System for Coordination of Recommendations to Reduce Resource Contention

Disclosed is a system and method for coordination of recommendations by a service to reduce resource contention.

Recommendation systems focus on giving individual consumers the best possible advice, based on a variety of information. But they do not coordinate recommendations across users to ensure an optimized use of resources. For example, if the system leads too many people to the same resource -- a specific restaurant for example -- that resource may become congested leading to a sub-optimal experience for many of the people.

A better solution is to keep track of what recommendations have been given to all users, and optionally the statistical probability that those recommendations will be accepted. This would allow the system to balance demand for limited resources. Recommendations are coordinated across users in order to provide an optimal balance of resource utilization and user satisfaction.

In the simplest case, this coordination occurs across users within the same system. In a more complex case, this coordination would occur across different systems. The system could even influence other types of systems such as sensor and control systems. Involving more systems increases the accuracy and effectiveness of the solution for all users.

The system and method keeps track of how many times a particular resource has been recommended, along with the statistical probability that each recommendation will be accepted. This probability is learned over time from the behavior of individual users. For example, if User A almost always accepts the recommendation of the system, then a recommendation to user A would carry more weight than a recommendation to user B who rarely accepts recommendations from the system. This statistical probability can be improved by differentiating between different types of recommendations. Note that this focus on individual user probabilities is different from predictive aspects of control systems, which are computed for patterns of use across the system -- for example, in context X, Y happens 85% of the time regardless of the profiles of the individuals involved.

By understanding how many people are likely to consume a given resource based on outstanding recommendations, the system can more effectively make future recommendations. For example, if a user requests a recommendation for a good Italian restaurant, the system may recommend the second best choice if the top choice has been recommended too many times already.

The invention works by executing the following steps:

A user requests a recommendation. This req...