Matches in SemOpenAlex for { <https://semopenalex.org/work/W3186355800> ?p ?o ?g. }
Showing items 1 to 86 of
86
with 100 items per page.
- W3186355800 abstract "Dynamic programming (DP) has a rich theoretical foundation and a broad range of applications, especially in the classic area of optimal control and the recent area of reinforcement learning (RL). Many optimal control problems can be solved as a single optimization problem, named one-shot optimization, or via a sequence of optimization problems using DP. However, the computation of their global optima often faces the NP-hardness issue due to the non-linearity of the dynamics and non-convexity of the cost, and thus only local optimal solutions may be obtained at best. Furthermore, in many cases arising in machine learning and model-free approaches, DP is the only viable choice, and therefore it is essential to understand when DP combined with a local search solver works. In this work, we introduce the notions of spurious local minimizers for the one-shot optimization and spurious local minimum policies for DP, and show that there is a deep connection between them. In particular, we prove that under mild conditions the DP method using local search can successfully solve the optimal control problem to global optimality if and only if the one-shot optimization is free of spurious solutions. This result paves the way to understand the performance of local search methods in optimal control and RL." @default.
- W3186355800 created "2021-08-02" @default.
- W3186355800 creator A5027576572 @default.
- W3186355800 creator A5042580848 @default.
- W3186355800 creator A5057565845 @default.
- W3186355800 date "2021-05-25" @default.
- W3186355800 modified "2023-09-24" @default.
- W3186355800 title "Analysis of Spurious Local Solutions of Optimal Control Problems: One-Shot Optimization Versus Dynamic Programming" @default.
- W3186355800 cites W1539331983 @default.
- W3186355800 cites W2098432798 @default.
- W3186355800 cites W2120087366 @default.
- W3186355800 cites W2121863487 @default.
- W3186355800 cites W2145339207 @default.
- W3186355800 cites W2163660424 @default.
- W3186355800 cites W2341171179 @default.
- W3186355800 cites W2899748887 @default.
- W3186355800 cites W2948432982 @default.
- W3186355800 cites W2963417959 @default.
- W3186355800 cites W2964225338 @default.
- W3186355800 cites W2970216099 @default.
- W3186355800 cites W2990006450 @default.
- W3186355800 cites W3090169104 @default.
- W3186355800 cites W3099560202 @default.
- W3186355800 cites W3123272904 @default.
- W3186355800 doi "https://doi.org/10.23919/acc50511.2021.9483238" @default.
- W3186355800 hasPublicationYear "2021" @default.
- W3186355800 type Work @default.
- W3186355800 sameAs 3186355800 @default.
- W3186355800 citedByCount "0" @default.
- W3186355800 crossrefType "proceedings-article" @default.
- W3186355800 hasAuthorship W3186355800A5027576572 @default.
- W3186355800 hasAuthorship W3186355800A5042580848 @default.
- W3186355800 hasAuthorship W3186355800A5057565845 @default.
- W3186355800 hasConcept C106159729 @default.
- W3186355800 hasConcept C119857082 @default.
- W3186355800 hasConcept C126255220 @default.
- W3186355800 hasConcept C135320971 @default.
- W3186355800 hasConcept C137836250 @default.
- W3186355800 hasConcept C141934464 @default.
- W3186355800 hasConcept C154945302 @default.
- W3186355800 hasConcept C162324750 @default.
- W3186355800 hasConcept C2778770139 @default.
- W3186355800 hasConcept C33923547 @default.
- W3186355800 hasConcept C37404715 @default.
- W3186355800 hasConcept C41008148 @default.
- W3186355800 hasConcept C72134830 @default.
- W3186355800 hasConcept C91575142 @default.
- W3186355800 hasConcept C97256817 @default.
- W3186355800 hasConcept C97541855 @default.
- W3186355800 hasConceptScore W3186355800C106159729 @default.
- W3186355800 hasConceptScore W3186355800C119857082 @default.
- W3186355800 hasConceptScore W3186355800C126255220 @default.
- W3186355800 hasConceptScore W3186355800C135320971 @default.
- W3186355800 hasConceptScore W3186355800C137836250 @default.
- W3186355800 hasConceptScore W3186355800C141934464 @default.
- W3186355800 hasConceptScore W3186355800C154945302 @default.
- W3186355800 hasConceptScore W3186355800C162324750 @default.
- W3186355800 hasConceptScore W3186355800C2778770139 @default.
- W3186355800 hasConceptScore W3186355800C33923547 @default.
- W3186355800 hasConceptScore W3186355800C37404715 @default.
- W3186355800 hasConceptScore W3186355800C41008148 @default.
- W3186355800 hasConceptScore W3186355800C72134830 @default.
- W3186355800 hasConceptScore W3186355800C91575142 @default.
- W3186355800 hasConceptScore W3186355800C97256817 @default.
- W3186355800 hasConceptScore W3186355800C97541855 @default.
- W3186355800 hasFunder F4320306076 @default.
- W3186355800 hasFunder F4320337345 @default.
- W3186355800 hasFunder F4320338279 @default.
- W3186355800 hasFunder F4320338281 @default.
- W3186355800 hasLocation W31863558001 @default.
- W3186355800 hasOpenAccess W3186355800 @default.
- W3186355800 hasPrimaryLocation W31863558001 @default.
- W3186355800 hasRelatedWork W1500457686 @default.
- W3186355800 hasRelatedWork W2003848856 @default.
- W3186355800 hasRelatedWork W200853570 @default.
- W3186355800 hasRelatedWork W2043434758 @default.
- W3186355800 hasRelatedWork W2126554300 @default.
- W3186355800 hasRelatedWork W2952680837 @default.
- W3186355800 hasRelatedWork W3132140784 @default.
- W3186355800 hasRelatedWork W3186355800 @default.
- W3186355800 hasRelatedWork W3197880104 @default.
- W3186355800 hasRelatedWork W4327643782 @default.
- W3186355800 isParatext "false" @default.
- W3186355800 isRetracted "false" @default.
- W3186355800 magId "3186355800" @default.
- W3186355800 workType "article" @default.