Particle swarm optimization method (PSO) and random search (Monte Carlo)

Presented implementations

This repository contains sequential and parallel implementation made with OpenMP. You can find our other implementations in the following repositories:

MPI: https://github.com/maciejpik/porr-pso-mpi
CUDA: https://github.com/maciejpik/porr-pso-cuda

Exemplary tasks the software optimizes

Task 1

$\min_{x} \left ( f(x)=\frac{1}{40}\sum_{i=1}^{n}(x_i^2) +1 - \prod_{i=1}^{n}cos(\frac{x_i}{i}) \right ), -40\leq x_i\leq 40, i=1,\dots,n$

Task 2

$\min_{x} \left ( f_1(x)=\sum_{i=1}^{n-1} \left [ 100(x_{i+1}-x_i^2)^2 + (1-x_i)^2) \right ] \right ), -40\leq x_i\leq 40$

2d sections of our tasks with the heat map

Task 1

Task 2

Exemplary results

Task 1 - PSO

Task 1 - Monte Carlo

Task 2 - PSO

Task 2 - Monte Carlo

Task 2

Generated with MPI Monte carlo version for Task 2 - Valley/Banana function

Trajectories

Monte carlo Task 1

PSO Task 1

Generated with MPI Monte carlo version for Task 1

Generated with MPI PSO version for Task 1

Plots with isolines

Task 1 - Isolines of cost function

Task 2 - Neighbourhood of global minimum of Rosenbrock's function

egi-megi / porr-pso

readme

Particle swarm optimization method (PSO) and random search (Monte Carlo)

Presented implementations

Exemplary tasks the software optimizes

Task 1

Task 2

2d sections of our tasks with the heat map

Task 1

Task 2

Exemplary results

Task 1 - PSO

Task 1 - Monte Carlo

Task 2 - PSO

Task 2 - Monte Carlo

Task 2

Generated with MPI Monte carlo version for Task 2 - Valley/Banana function

Trajectories

Monte carlo Task 1

PSO Task 1

Generated with MPI Monte carlo version for Task 1

Generated with MPI PSO version for Task 1

Plots with isolines

Task 1 - Isolines of cost function

Task 2 - Neighbourhood of global minimum of Rosenbrock's function