SEARCH, blackbox optimization, and sample complexity
- Los Alamos National Lab., NM (United States). Computational Science Methods Div.
- Univ. of Illinois, Urbana, IL (United States). Dept. of General Engineering
The SEARCH (Search Envisioned As Relation and Class Hierarchizing) framework developed elsewhere (Kargupta, 1995; Kargupta and Goldberg, 1995) offered an alternate perspective toward blackbox optimization -- optimization in presence of little domain knowledge. The SEARCH framework investigates the conditions essential for transcending the limits of random enumerative search using a framework developed in terms of relations, classes and partial ordering. This paper presents a summary of some of the main results of that work. A closed form bound on the sample complexity in terms of the cardinality of the relation space, class space, desired quality of the solution and the reliability is presented. This also leads to the identification of the class of order-k delineable problems that can be solved in polynomial sample complexity. These results are applicable to any blackbox search algorithms, including evolutionary optimization techniques.
- Research Organization:
- Los Alamos National Lab. (LANL), Los Alamos, NM (United States)
- Sponsoring Organization:
- Department of the Air Force, Washington, DC (United States)
- DOE Contract Number:
- W-7405-ENG-36
- OSTI ID:
- 251407
- Report Number(s):
- LA-UR-96-430; CONF-960820-4; ON: DE96008114; CNN: Grant F49620-94-1-0103; TRN: AHC29614%%135
- Resource Relation:
- Conference: Foundations of genetic algorithms, San Diego, CA (United States), 3-6 Aug 1996; Other Information: PBD: [1996]
- Country of Publication:
- United States
- Language:
- English
Similar Records
Blackbox and non-blackbox optimization: A common perspective
SEARCH and a computational perspective of evolution