Evolutionary strategy-python Implementation

esindividual.py

1 ImportNumPy as NP2 Importobjfunction3 4 5 classesindividual:6 7     " "8 individual of evolutionary strategy9     " "Ten  One     def __init__(self, Vardim, bound): A         " " - Vardim:dimension of Variables - Bound:boundaries of Variables the         " " -Self.vardim =Vardim -Self.bound =bound -Self.fitness =0. +Self.trials =0 -  +     defGenerate (self): A         " " at generate a random chromsome for evolutionary strategy -         " " -Len =Self.vardim -Rnd = Np.random.random (size=len) -Self.chrom =Np.zeros (len) -          forIinchxrange (0, Len): inSelf.chrom[i] = self.bound[0, I] +  -(Self.bound[1, I]-self.bound[0, I]) *Rnd[i] to  +     defcalculatefitness (self): -         " " the calculate the fitness of the Chromsome *         " " $Self.fitness =Objfunction.griefunc (Panax NotoginsengSelf.vardim, Self.chrom, Self.bound)

es.py

1 ImportNumPy as NP2  fromEsindividualImportesindividual3 ImportRandom4 ImportCopy5 ImportMatplotlib.pyplot as Plt6 7 8 classEvolutionarystrategy:9 Ten     " " One The class for evolutionary strategy A     " " -  -     def __init__(self, sizepop, Vardim, bound, Maxgen, params): the         " " - sizepop:population Sizepop - Vardim:dimension of Variables - Bound:boundaries of Variables + maxgen:termination Condition - params:algorithm Required parameters, it is a list which is consisting Of[delta_max, Delta_min] +         " " ASelf.sizepop =Sizepop atSelf.vardim =Vardim -Self.bound =bound -Self. Maxgen =Maxgen -Self.params =params -Self.population = [] -Self.fitness =Np.zeros (Self.sizepop) inSelf.trace = Np.zeros (self. Maxgen, 2)) -  to     defInitialize (self): +         " " - Initialize the population of ES the         " " *          forIinchxrange (0, Self.sizepop): $IND =esindividual (Self.vardim, Self.bound)Panax Notoginseng ind.generate () - Self.population.append (Ind) the  +     defEvaluation (self): A         " " the Evaluation The fitness of the population +         " " -          forIinchxrange (0, Self.sizepop): $ self.population[i].calculatefitness () $Self.fitness[i] =self.population[i].fitness -  -     defSolve (self): the         " " - The evolution process of the evolutionary strategyWuyi         " " theSELF.T =0 - self.initialize () Wu self.evaluation () -Bestindex =Np.argmax (self.fitness) AboutSelf.best =copy.deepcopy (Self.population[bestindex]) $          whileSELF.T <Self . Maxgen: -SELF.T + = 1 -Tmppop =self.mutation () - self.selection (Tmppop) ABest =Np.max (self.fitness) +Bestindex =Np.argmax (self.fitness) the             ifBest >self.best.fitness: -Self.best =copy.deepcopy (Self.population[bestindex]) $  theSelf.avefitness =Np.mean (self.fitness) theSelf.trace[self.t-1, 0] =  the(1-self.best.fitness)/self.best.fitness theSelf.trace[self.t-1, 1] = (1-self.avefitness)/self.avefitness -             Print("Generation%d:optimal function value is:%f, average function value is%f"% ( inSELF.T, self.trace[self.t-1, 0], self.trace[self.t-1, 1])) the         Print("Optimal function value is:%f;"% self.trace[self.t-1, 0]) the         Print "Optimal solution is:" About         PrintSelf.best.chrom the Self.printresult () the  the     defmutation (self): +         " " - mutate the population by a random normal distribution the         " "BayiTmppop = [] the          forIinchxrange (0, Self.sizepop): theIND =copy.deepcopy (Self.population[i]) -Delta = self.params[0] + self.t *  -(Self.params[1]-self.params[0])/Self . Maxgen theInd.chrom + = Np.random.normal (0.0, Delta, Self.vardim) the              forKinchxrange (0, Self.vardim): the                 ifIND.CHROM[K] <self.bound[0, K]: theIND.CHROM[K] =self.bound[0, K] -                 ifInd.chrom[k] > Self.bound[1, K]: theInd.chrom[k] = self.bound[1, K] the ind.calculatefitness () the Tmppop.append (Ind)94         returnTmppop the  the     defselection (self, tmppop): the         " "98 Update the population About         " " -          forIinchxrange (0, Self.sizepop):101             ifSelf.fitness[i] <tmppop[i].fitness:102Self.population[i] =Tmppop[i]103Self.fitness[i] =tmppop[i].fitness104  the     defPrintresult (self):106         " "107 plot The result of evolutionary strategy108         " "109x =np.arange (0, self. Maxgen) theY1 =self.trace[:, 0]111y2 = self.trace[:, 1] thePlt.plot (x, y1,'R', label='Optimal Value')113Plt.plot (x, Y2,'g', label='Average value') thePlt.xlabel ("Iteration") thePlt.ylabel ("function Value") thePlt.title ("evolutionary strategy for function optimization")117 plt.legend ()118Plt.show ()

To run the program:

1 if __name__ " __main__ " : 2 3     bound = Np.tile ([[ -600], [+]],    4     es = es (1, bound, +, [])
   
    5     es.solve ()
   

Objfunction See simple genetic algorithm-python realization.

Evolutionary strategy-python Implementation