Tree-Type Decision Model for Advertisement Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current advertisement selection systems face challenges in efficiently maximizing revenue for publishers by optimizing advertisement placement across multiple contracts and complex business models, both at the publisher Headend and resource-constrained end-user devices, particularly when dealing with large numbers of advertisements and flexible targeting criteria.
Innovation Solution
A system that performs global optimization at the publisher Headend and local optimization at end-user devices using a tree-type decision model, which evaluates targeting criteria to select and prioritize advertisements, optimizing the model based on static and transient user data to ensure compliance with advertiser requirements and reduce computational load.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If global optimization is performed at the publisher Headend for all advertisement placement decisions, then revenue maximization is improved, but the system cannot adapt to individual client characteristics and playback conditions
Solution Approach 1:
The patent divides the advertisement decision-making system into two segments: a publisher Headend that performs global optimization for revenue maximization, and client-side agents that perform local optimization adapted to individual client characteristics, playback conditions, and device constraints. This segmentation allows both global revenue goals and local adaptability to be achieved simultaneously.
2Adaptability or versatility
If local optimization is performed at end-user devices with access to hundreds or thousands of advertisements, then adaptability to client characteristics is improved, but the computational load becomes too great for resource-constrained devices
Solution Approach 1:
The patent applies preliminary action by having the publisher Headend pre-process and filter the advertisement catalog before distribution to clients. The global optimization at the Headend pre-determines which advertisements are relevant for each client based on available information, so that clients receive a reduced, manageable subset of advertisements rather than having to evaluate hundreds or thousands of options locally.
3Adaptability or versatility
If the advertisement catalog is distributed to all clients, then local adaptability is improved, but the storage requirements at resource-constrained devices increase
Solution Approach 1:
The patent applies local quality by distributing different subsets of advertisements to different clients based on their specific characteristics, content subscriptions, and playback conditions. Each client receives a customized, reduced catalog tailored to its local needs rather than a complete universal catalog, optimizing storage efficiency while maintaining local adaptability.
Data Source
AI summary
An end-user rendering system including an advertisement database to receive advertisements, and store the advertisements therein, a state database to store information, a decision model optimization module to receive a tree-type decision model and optimize the tree-type decision model based on at least some of the information stored in the state database, an advertisement decision module to evaluate the optimized tree-type decision model and select an advertising campaign, the selected advertising campaign having at least one advertisement, and a rendering module to render the at least one advertisement of the selected advertising campaign. Related apparatus and methods are also described.