my_optimization

import os
import ipdb
import numpy as np
from random import randint

# My Hill Climbing
def my_hillclimbing(fun, x0, bounds, maxiter, step, args):
     x = x0
     cost = fun(x,*args)
     best_cost = cost
     best_x = x

     xmin = []
     xmax = []
     for i in range(0,len(bounds)):
          xmin.append(bounds[i][0])
          xmax.append(bounds[i][1])

     iteration = 0
     for i in range(0,maxiter):  #stop criterion
     #while (best_cost > 0.3 or iteration > maxiter):
          first = 1
          new_x = x.copy()
          while (first or (not (all(new_x>=xmin) and all(new_x<=xmax)))):
               new_x = x.copy()
               first = 0
               idx = randint(0,len(x)-1)
               signal = randint(0,1)
               new_x[idx] = new_x[idx]+(int(signal)*step)-(int(not(signal))*step)
               #print("trying new_x[%d] = %f" %(idx,new_x[idx]))

          new_cost = fun(new_x,*args)
          print ("new_x: %s \t new_cost: %s" % (new_x, new_cost))
          if new_cost <= best_cost:
               best_cost = new_cost
               best_x = new_x.copy()

          if (new_cost < cost):
               cost = new_cost
               x = new_x.copy()

          iteration += 1

     class list_results(object):
          def __init__(self):
               self.x = best_x
               self.cost = best_cost

     res = list_results()

     return (res);