Information Processing Device for Online Submodular Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for online submodular optimization, such as those used in advertising and sales, struggle to derive effective subsets when a fixed strategy is not applicable, leading to inefficiencies in minimizing or maximizing cumulative values over rounds.
Innovation Solution
An information processing apparatus that sets objective functions for each round and derives subset sequences with controlled regret values, using Hamming distance constraints to ensure effective subset selection even when a fixed strategy is ineffective.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a fixed strategy to select the same subset in all rounds is used, then the expected value of regret is minimized to not more than O((nT)1/2), but useful subsets cannot be derived for online submodular optimization problems for which a fixed strategy is not effective
Solution Approach 1:
The patent applies dynamics by transitioning from a static fixed strategy to a dynamic adaptive strategy. The algorithm maintains a set of candidate subsets and dynamically selects among them based on feedback from objective function evaluations in each round, allowing the strategy to adapt to changing problem conditions while controlling regret growth.
Solution Approach 2:
The patent employs feedback mechanisms where the algorithm evaluates the objective function for selected subsets in each round and uses this feedback to update its strategy. This feedback loop enables the system to learn from past performance and adjust future selections, thereby deriving useful subsets for problems where fixed strategies fail.
2Ease of manufacture
If a fixed strategy is used for subset selection, then the algorithm is simple to implement, but it cannot handle online submodular optimization problems where the optimal strategy changes over time
Solution Approach 1:
The patent segments the algorithm into distinct functional components: an evaluation phase where the objective function is computed for candidate subsets, and a selection phase where the best subset is chosen based on current feedback. This segmentation maintains organizational simplicity while enabling adaptive optimization effectiveness.
Solution Approach 2:
The patent applies preliminary action by pre-establishing a set of candidate subsets before the optimization process begins. This preliminary preparation simplifies the ongoing algorithm execution while enabling effective adaptation to time-varying problems through selective evaluation and update mechanisms.
Data Source
AI summary
In order to enable derivation of useful subsets X1, X2, . . . , XT also for an online submodular optimization problem for which a fixed strategy is not effective, an information processing apparatus (1) includes: an objective function setting unit (11) that sets, as an objective function ft in each round t∈[T], a submodular function on a power set 2S of a set S consisting of n elements; and a subset sequence derivation unit (12) that derives a subset sequence X1, X2, . . . , XT∈2S in which an expected value of regret Σt∈[T]ft(Xt)−Σt∈[T]ft(Xt*) with respect to any benchmark X1*, X2*, . . . , Xt*∈2S satisfying Σt∈[T−1]dH(Xt*, Xt+1*) is not more than an upper limit Max (n,T,V).