code

PSO Implementation (in Ruby)

def objective_function(vector) return vector.inject(0.0) {|sum, x| sum + (x ** 2.0)}end
def random_vector(minmax) return Array.new(minmax.size) do |i| minmax[i][0] + ((minmax[i][1] - minmax[i][0]) * rand()) endend
def create_particle(search_space, vel_space) particle = {} particle[:position] = random_vector(search_space) particle[:cost] = objective_function(particle[:position]) particle[:b_position] = Array.new(particle[:position]) particle[:b_cost] = particle[:cost] particle[:velocity] = random_vector(vel_space) return particleend
def get_global_best(population, current_best=nil) population.sort!{|x,y| x[:cost] <=> y[:cost]} best = population.first if current_best.nil? or best[:cost] <= current_best[:cost] current_best = {} current_best[:position] = Array.new(best[:position]) current_best[:cost] = best[:cost] end return current_bestend
def update_velocity(particle, gbest, max_v, c1, c2) particle[:velocity].each_with_index do |v,i| v1 = c1 * rand() * (particle[:b_position][i] - particle[:position][i]) v2 = c2 * rand() * (gbest[:position][i] - particle[:position][i]) particle[:velocity][i] = v + v1 + v2 particle[:velocity][i] = max_v if particle[:velocity][i] > max_v particle[:velocity][i] = -max_v if particle[:velocity][i] < -max_v endend
def update_position(part, bounds) part[:position].each_with_index do |v,i| part[:position][i] = v + part[:velocity][i] if part[:position][i] > bounds[i][1] part[:position][i]=bounds[i][1]-(part[:position][i]-bounds[i][1]).abs part[:velocity][i] *= -1.0 elsif part[:position][i] < bounds[i][0] part[:position][i]=bounds[i][0]+(part[:position][i]-bounds[i][0]).abs part[:velocity][i] *= -1.0 end endend
def update_best_position(particle) return if particle[:cost] > particle[:b_cost] particle[:b_cost] = particle[:cost] particle[:b_position] = Array.new(particle[:position])end
def search(max_gens, search_space, vel_space, pop_size, max_vel, c1, c2) pop = Array.new(pop_size) {create_particle(search_space, vel_space)} gbest = get_global_best(pop) max_gens.times do |gen| pop.each do |particle| update_velocity(particle, gbest, max_vel, c1, c2) update_position(particle, search_space) particle[:cost] = objective_function(particle[:position]) update_best_position(particle) end gbest = get_global_best(pop, gbest) puts " > gen #{gen+1}, fitness=#{gbest[:cost]}" end return gbestend
# problem configuration problem_size = 2 search_space = Array.new(problem_size) {|i| [-5, 5]} # algorithm configuration vel_space = Array.new(problem_size) {|i| [-1, 1]} max_gens = 100 pop_size = 50 max_vel = 100.0 c1, c2 = 2.0, 2.0 # execute the algorithm best = search(max_gens, search_space, vel_space, pop_size, max_vel, c1,c2) puts "done! Solution: f=#{best[:cost]}, s=#{best[:position].inspect}"

Output

 > gen 1, fitness=0.10732377938785477

 > gen 2, fitness=0.008043798087950995

 > gen 3, fitness=0.008043798087950995

 > gen 4, fitness=0.008043798087950995

 > gen 5, fitness=0.008043798087950995

 > gen 6, fitness=0.008043798087950995

 > gen 7, fitness=0.008043798087950995

 > gen 8, fitness=0.008043798087950995

 > gen 9, fitness=0.008043798087950995

 > gen 10, fitness=0.008043798087950995

 > gen 11, fitness=0.008043798087950995

 > gen 12, fitness=0.008043798087950995

 > gen 13, fitness=0.008043798087950995

 > gen 14, fitness=0.008043798087950995

 > gen 15, fitness=0.008043798087950995

 > gen 16, fitness=0.008043798087950995

 > gen 17, fitness=0.008043798087950995

 > gen 18, fitness=0.008043798087950995

 > gen 19, fitness=0.008043798087950995

 > gen 20, fitness=0.008043798087950995

 > gen 21, fitness=0.008043798087950995

 > gen 22, fitness=0.008043798087950995

 > gen 23, fitness=0.008043798087950995

 > gen 24, fitness=0.008043798087950995

 > gen 25, fitness=0.008043798087950995

 > gen 26, fitness=0.008043798087950995

 > gen 27, fitness=0.008043798087950995

 > gen 28, fitness=0.008043798087950995

 > gen 29, fitness=0.008043798087950995

 > gen 30, fitness=0.008043798087950995

 > gen 31, fitness=0.008043798087950995

 > gen 32, fitness=0.008043798087950995

 > gen 33, fitness=0.008043798087950995

 > gen 34, fitness=0.008043798087950995

 > gen 35, fitness=0.008043798087950995

 > gen 36, fitness=0.008043798087950995

 > gen 37, fitness=0.008043798087950995

 > gen 38, fitness=0.008040353298287386

 > gen 39, fitness=0.008040353298287386

 > gen 40, fitness=0.008040353298287386

 > gen 41, fitness=0.0033875006805440437

 > gen 42, fitness=0.0033875006805440437

 > gen 43, fitness=0.0033875006805440437

 > gen 44, fitness=0.0033875006805440437

 > gen 45, fitness=0.0033875006805440437

 > gen 46, fitness=0.0033875006805440437

 > gen 47, fitness=0.0033875006805440437

 > gen 48, fitness=0.0033875006805440437

 > gen 49, fitness=0.0033875006805440437

 > gen 50, fitness=0.0033875006805440437

 > gen 51, fitness=0.0033875006805440437

 > gen 52, fitness=0.0033875006805440437

 > gen 53, fitness=0.0033875006805440437

 > gen 54, fitness=0.0033875006805440437

 > gen 55, fitness=0.0033875006805440437

 > gen 56, fitness=0.0033875006805440437

 > gen 57, fitness=0.0033875006805440437

 > gen 58, fitness=0.0033875006805440437

 > gen 59, fitness=0.0033875006805440437

 > gen 60, fitness=0.0033875006805440437

 > gen 61, fitness=0.0033875006805440437

 > gen 62, fitness=0.0033875006805440437

 > gen 63, fitness=0.0033875006805440437

 > gen 64, fitness=0.0033875006805440437

 > gen 65, fitness=0.0033875006805440437

 > gen 66, fitness=0.0033875006805440437

 > gen 67, fitness=0.0033875006805440437

 > gen 68, fitness=0.0033875006805440437

 > gen 69, fitness=0.0033875006805440437

 > gen 70, fitness=0.0033875006805440437

 > gen 71, fitness=0.0033875006805440437

 > gen 72, fitness=0.0033875006805440437

 > gen 73, fitness=0.0033875006805440437

 > gen 74, fitness=0.0033875006805440437

 > gen 75, fitness=0.0033875006805440437

 > gen 76, fitness=0.0033875006805440437

 > gen 77, fitness=0.0033875006805440437

 > gen 78, fitness=0.0033875006805440437

 > gen 79, fitness=0.0033875006805440437

 > gen 80, fitness=0.0033875006805440437

 > gen 81, fitness=0.0033875006805440437

 > gen 82, fitness=0.0033875006805440437

 > gen 83, fitness=0.0033875006805440437

 > gen 84, fitness=0.0033875006805440437

 > gen 85, fitness=0.0033875006805440437

 > gen 86, fitness=0.0033875006805440437

 > gen 87, fitness=0.0033875006805440437

 > gen 88, fitness=0.0033875006805440437

 > gen 89, fitness=0.0033875006805440437

 > gen 90, fitness=0.0033875006805440437

 > gen 91, fitness=0.0033875006805440437

 > gen 92, fitness=0.0033875006805440437

 > gen 93, fitness=0.0033875006805440437

 > gen 94, fitness=0.0033875006805440437

 > gen 95, fitness=0.0033875006805440437

 > gen 96, fitness=0.0033875006805440437

 > gen 97, fitness=0.0033875006805440437

 > gen 98, fitness=0.0033875006805440437

 > gen 99, fitness=0.0033875006805440437

 > gen 100, fitness=0.0033875006805440437

done! Solution: f=0.0033875006805440437, s=[0.05632502572719389, -0.014662610864883185]

Google Sites

Report abuse