Introduction

We concerns global optimization (GO) problems, in which one wants to get the globally best solution of an objective function f(x) on a given domain, where f might be multi-modal, non-differentiable, discontinuous, or even worse black-box type.      Difficulties:     1. Derivative-based methods are easily get trapped into local optimum.     2. Derivative of objective function is unavailable, or hard tocompute, or         unreliable (numerically unstable, e.g., in the presence of noise).             So derivative-free methods, including evolutionary algorithms, are preferred.       Some efficient evolutionary algorithms for GO:     1. Real-coded genetic algorithm (RGA)     2. Genetic programming (GP)     3. Particle swarm optimization (PSO)     4. Differential evolution (DE)     5. Low dimensional simplex evolution (LDSE) To discuss with the colleagues who work in global optimization field,  please visit Google group on global optimization.

LDSE Algorithm

Low dimensional simplex evolution (LDSE) is a kind of real-coded evolutionary algorithm for global optimization. It generates new individuals in Nelder-Mead way, but the simplex is low and variable dimensional,  and the simplex operators are applied conditionally and selectively on low dimensional subspaces. The individuals survive by the rule of natural selection in the framework of try-try-struggle. Recently we released a demo product of LDSE algorithm: LDSE for Linux with GTK+ 2.x LDSE for Mac OS PowerPC LDSE for Mac OS Universal Binary LDSE for Mac OS intel LDSE for Windows 2000 or later Examples of using LDSE To discuss with the authors and users of LDSE, please visit Google group of this site. ===Earlier version=== The most simple form of LDSE is Triangle Evolution (TE). As a demo product of TE, TEOptimizer is an easy-to-use tool for global optimization, which is attached below. File name:                TEOptimizer.rar System requirement: Windows XP or later