Ko-Hsin Liang, Xin Yao, and Charles Newton. Combining landscape approximation and local search in global optimization. In Proc. of the 1999.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....local optima of the original multi modal tness function without changing the global optimum and its location. A Gaussian kernel has been used to realize coarse to ne smoothing of the original multi modal function [84] Approximation for smoothing multi modal functions has also been reported in [45, 46], where global polynomial models are used instead of Gaussian kernel functions. Note that it is generally dicult to build an approximate model that has the same global optimum on the same location. Therefore, the coarse to ne modeling approach seems to be more realistic. 3 Levels of ....

K.-H. Liang, X. Yao, and C. Newton. Combining landscape approximation and local search in global optimization. In 1999 Congress on Evolutionary Computation, pages 1514-1520, 1999.

....optimum and its location. The approximated and smoothened landscape is often much easier to search than the original one. Our previous work on the combination of landscape approximation and local search (LALS) with evolutionary algorithms has shown some very promising results along this direction [11]. This paper further extends our previous method and proposes a novel evolutionary algorithm with n dimensional approximation (EANA) The n dimensional approximation can be regarded as a recombination operator in our new algorithm which integrates evolutionary and local search. Numerical ....

.... successful examples, for example the evolutionary algorithms with kriging approximation [15, 16] the genetic algorithms with local search [23, 9] the crossover operators with approximation concepts [19, 1] and the evolutionary algorithm with both landscape approximation and local search (LALS) [11]. The LALS algorithm [11] demonstrated high reliability in nding the global optimum of the benchmark multimodal problems. In this paper, we propose a novel evolutionary algorithm with n dimensional approximation (EANA) which shares the same motivation as LALS, but remedies some of the drawbacks ....

[Article contains additional citation context not shown here]

K.-H. Liang, X. Yao, and C. Newton. Combining landscape approximation and local search in global optimization. In Proceedings of the 1999 Congress on Evolutionary Computation, volume 2, pages 1514-1520, Piscataway, NJ, 1999. IEEE Press.

.... applications, for example the evolutionary algorithms with kriging approximation [23, 24] the evolutionary algorithms with local search [32, 13, 7] the crossover operators with approximation concepts [28, 2] and the evolutionary algorithm with both landscape approximation and local search (LALS) [17]. The LALS algorithm demonstrated high reliability in finding the global optimum of the benchmark multimodal problems. In this paper, we propose a new algorithm using landscape approximation and hybrid evolutionary and local search. We list several algorithm design principles. Following the basic ....

....method. A basic principle of applying approximation techniques is that the approximated function can be easily calculated to get the approximated minimum location. A quadratic polynomial with least squares approximation is a good example and has been applied in some optimization methods [33, 22, 10, 17]. The DACE model based on Bayesian statistics can give more precise approximation about the landscape, however other auxiliary global optimization algorithms are needed to obtain the approximated minimum from the approximation function. Different approximation methods also demand different ....

[Article contains additional citation context not shown here]

K.-H. Liang, X. Yao, and C. Newton. Combining landscape approximation and local search in global optimization. In Proceedings of the 1999 Congress on Evolutionary Computation, volume 2, pages 1514--1520, Piscataway, NJ, 1999. IEEE Press.

No context found.

Ko-Hsin Liang, Xin Yao, and Charles Newton. Combining landscape approximation and local search in global optimization. In Proc. of the 1999.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC