Search Results for "differential evolution"

The Differential Evolution, introduced in 1995 by Storn and Price, considers the population, that is divided into branches, one per computational node. The Differential Evolution Entirely Parallel method takes into account the individual age, that is defined as the number of iterations the individual survived without changes. The introduced improvements are: (I) allow several oldest individuals to be overwritten by the same number of best ones in the population, (II) new selection rule uses several objective functions in offspring evaluation. ...

The ASCO project aims to bring circuit optimization capabilities to existing SPICE simulators using a high-performance parallel differential evolution (DE) optimization algorithm. It supports Eldo, HSPICE, LTspice, Spectre, and Qucs.

popt4jlib is an open-source parallel optimization library for the Java programming language supporting both shared memory and distributed message passing models. Implements a number of meta-heuristic algorithms for Non-Linear Programming, including Genetic Algorithms, Differential Evolution, Evolutionary Algorithms, Simulated Annealing, Particle Swarm Optimization, Firefly Algorithm, Monte-Carlo Search, Local Search algorithms, Gradient-Descent-based algorithms, as well as some well-known network flow and other graph algorithms. A fast parallel implementation of the network simplex method, and some full-fledged parallel/distributed MIP solvers will be added in the next version. ...

Opt4J is an open source Java-based framework for evolutionary computation. It contains a set of (multi-objective) optimization algorithms such as evolutionary algorithms (including SPEA2 and NSGA2), differential evolution, particle swarm optimization, and simulated annealing. The benchmarks that are included comprise ZDT, DTLZ, WFG, and the knapsack problem. The goal of Opt4J is to simplify the evolutionary optimization of user-defined problems as well as the implementation of arbitrary meta-heuristic optimization algorithms. For this purpose, Opt4J relies on a module-based implementation and offers a graphical user interface for the configuration as well as a visualization of the optimization process.

The MOEA Framework is a free and open source Java library for developing and experimenting with multiobjective evolutionary algorithms (MOEAs) and other general-purpose multiobjective optimization algorithms. The MOEA Framework supports genetic algorithms, differential evolution, particle swarm optimization, genetic programming, grammatical evolution, and more. A number of algorithms are provided out-of-the-box, including NSGA-II, NSGA-III, ε-MOEA, GDE3 and MOEA/D. In addition, the MOEA Framework provides the tools necessary to rapidly design, develop, execute and statistically test optimization algorithms.

Keywords: -simultaneously fit several SAXS and SANS data sets with polydisperse (Schultz-Zimm or Gaussian distribution f(R)) spherical core-shell-shell nanoparticles -analytical expressions are used for from factor F(Q) and its integral over f(R), no numerical integration required -absolute units -Mathematica is required via console (MathKernel) -Mathematica's local and global optimizers (simulated annealing, differential evolution, Nelder-Mead, ...) can be used -range for fit parameters and further constraints between fit parameters are possible -Monodisperse(!) hard sphere structure factor can be used, too -long computation times (depending on problem size and amount of constraints) from hours to a few days are possible -non-parallelized code

Keywords (XNDiff): -SAXS -SANS -absolute units -core (double)shell crystalline nanoparticles -with a parallelepidal shape -particle assemblies -powder and ensemble average -C/C++ -Unix -OpenMP -HPC Cluster Keywords (BatchMultiFit): -simultaneous fits for several SAXS and SANS curves with simulation data from XNDiff -SANS data can be smeared with dq values from experimental data sets or analytical functions -Mathematica console -local and global optimizers (simulated annealing, differential evolution, Nelder-Mead, ...) can be used -range for fit parameters and further constraints between fit parameters -parallelized (typ. 4-8 threads) TODO (BatchMultiFit): -read and use errorbars from experimental data sets -allow different q-ranges for different data sets in the fits -rewrite and test in Python using e.g. the lmfit module: https://pypi.python.org/pypi/lmfit/ to get rid of Mathematica and to run it on HPC clusters

This is a GUI, Java implementation of the Ant Colony Optimisation/Particle Swarm Optimisation (PSO/ACO2) rule induction algorithm. This project was inspired by Ant-Miner, but handles continuous attributes using PSO or now Differential Evolution.