Local Update of Support Vector Machine Decision Boundaries

Abstract

This paper presents a new adaptive sampling technique for the construction of locally refined explicit decision functions. The decision functions can be used for both deterministic and probabilistic optimization, and may represent a constraint or a limit-state function. In particular, the focus of this paper is on reliability-based design optimization (RBDO). Instead of approximating the responses, the method is based on explicit design space decomposition (EDSD), in which an explicit boundary separating distinct regions in the design space is constructed. A statistical learning tool known as support vector machine (SVM) is used to construct the boundaries. A major advantage of using an EDSD-based method lies in its ability to handle discontinuous responses. A separate adaptive sampling scheme for calculating the probability of failure is also developed, which is used within the RBDO process. The update methodology is validated through several test examples with analytical decision functions. I.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 10 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.