Title :
Randomized Pattern Search
Author :
Lawrence, J.P. ; Steiglitz, Kenneth
Author_Institution :
Part-time student in the Department of Electrical Engineering, University of Maryland, College Park, Md.
fDate :
4/1/1972 12:00:00 AM
Abstract :
A random search technique for function minimization is proposed that incorporates the step-size and direction adaptivity of Hooke and Jeeves´ [1] pattern search. Experimental results for a variety of functions indicate that the random pattern search is more effective than the corresponding deterministic method for a class of problems with hard constraints.
Keywords :
Acceleration; Algorithms; Constraint optimization; Flowcharts; Minimization methods; Performance evaluation; Reluctance generators; Search methods; Virtual colonoscopy; Direct search; optimization algorithms; random search;
Journal_Title :
Computers, IEEE Transactions on
DOI :
10.1109/TC.1972.5008979
