Online Continuous DR-Submodular Maximization with Long-Term Budget Constraints

In this paper, we study a class of online optimization problems with long-term budget constraints where the objective functions are not necessarily concave (nor convex) but they instead satisfy the Diminishing Returns (DR) property. Specifically, a sequence of monotone DR-submodular objective functions and monotone linear budget functions arrive over time and assuming a total targeted budget , the goal is to choose points at each time , without knowing and on that step, to achieve sub-linear regret bound while the total budget violation is sub-linear as well. Prior work has shown that achieving sub-linear regret is impossible if the budget functions are chosen adversarially. Therefore, we modify the notion of regret by comparing the agent against a -approximation to the best fixed decision in hindsight which satisfies the budget constraint proportionally over any window of length . We propose the Online Saddle Point Hybrid Gradient (OSPHG) algorithm to solve this class of online problems. For , we recover the aforementioned impossibility result. However, when , we show that it is possible to obtain sub-linear bounds for both the -regret and the total budget violation.
View on arXiv