def optimize(**kwargs): min = kwargs['min'] max = kwargs['max'] optimizer = kwargs['optimizer'] del kwargs['min'] del kwargs['max'] del kwargs['optimizer'] return hve(optimizer(**kwargs), min, max, 100000)
def optimize(**kwargs): min=kwargs['min'] max=kwargs['max'] optimizer=kwargs['optimizer'] del kwargs['min'] del kwargs['max'] del kwargs['optimizer'] return hve(optimizer(**kwargs),min,max,100000)
max=max, Model=model, decnum=decnum, objnum=objnum, the_seed=-1, candidates=[100, 10, 1000], generations=1000, mutation_rate=[0.05, 0.01, 0.5], lifes=[5, 1, 10]) del kw['optimizer'] del kw['min'] del kw['max'] for the_seed in xrange(20): pf1 = GeneticAlgorithm(Model=model, decnum=decnum, objnum=objnum, the_seed=the_seed, candidates=candidates, generations=generations, mutation_rate=mutation_rate, lifes=lifes) pf2 = GeneticAlgorithm(**kw) #min,max=upminmax(min,max,pf) hv[model.__name__][objnum][decnum]['untuned'].append( hve(pf1, min, max, 100000)) hv[model.__name__][objnum][decnum]['tuned'].append( hve(pf2, min, max, 100000)) with open('./data/hypervolumn.pickle', 'wb') as handle: pickle.dump(hv, handle)
objs=[2,4,6,8] decs=[10,20,40] hv={} for model in models: hv[model.__name__]={} for objnum in objs: hv[model.__name__][objnum]={} for decnum in decs: hv[model.__name__][objnum][decnum]={'untuned':[],'tuned':[]} min,max=init(model,decnum=decnum,objnum=objnum,num=100000) kw=differential_evolution(Tunee,optimizer=GeneticAlgorithm,min=min,max=max,Model=model,decnum=decnum, objnum=objnum,the_seed=-1,candidates=[100,10,1000],generations=1000, mutation_rate=[0.05,0.01,0.5],lifes=[5,1,10]) del kw['optimizer'] del kw['min'] del kw['max'] for the_seed in xrange(20): pf1=GeneticAlgorithm(Model=model,decnum=decnum,objnum=objnum,the_seed=the_seed,candidates=candidates, generations=generations,mutation_rate=mutation_rate,lifes=lifes) pf2=GeneticAlgorithm(**kw) #min,max=upminmax(min,max,pf) hv[model.__name__][objnum][decnum]['untuned'].append(hve(pf1,min,max,100000)) hv[model.__name__][objnum][decnum]['tuned'].append(hve(pf2,min,max,100000)) with open('./data/hypervolumn.pickle', 'wb') as handle: pickle.dump(hv, handle)
models = [DTLZ1, DTLZ3, DTLZ5, DTLZ7] objs = [2, 4, 6, 8] decs = [10, 20, 40] hv = {} for model in models: hv[model.__name__] = {} for objnum in objs: hv[model.__name__][objnum] = {} for decnum in decs: hv[model.__name__][objnum][decnum] = [] min, max = init(model, decnum=decnum, objnum=objnum, num=100000) for the_seed in xrange(20): pf = GeneticAlgorithm( model, decnum=decnum, objnum=objnum, the_seed=the_seed, candidates=candidates, generations=generations, mutation_rate=mutation_rate, lifes=lifes, ) # min,max=upminmax(min,max,pf) hv[model.__name__][objnum][decnum].append(hve(pf, min, max, 100000)) with open("hypervolumn.pickle", "wb") as handle: pickle.dump(hv, handle) set_trace() with open("./data/hypervolumn.pickle", "wb") as handle: pickle.dump(hv, handle)
decs = [10, 20, 40] hv = {} for model in models: hv[model.__name__] = {} for objnum in objs: hv[model.__name__][objnum] = {} for decnum in decs: hv[model.__name__][objnum][decnum] = [] min, max = init(model, decnum=decnum, objnum=objnum, num=100000) for the_seed in xrange(20): pf = GeneticAlgorithm(model, decnum=decnum, objnum=objnum, the_seed=the_seed, candidates=candidates, generations=generations, mutation_rate=mutation_rate, lifes=lifes) #min,max=upminmax(min,max,pf) hv[model.__name__][objnum][decnum].append( hve(pf, min, max, 100000)) with open('hypervolumn.pickle', 'wb') as handle: pickle.dump(hv, handle) set_trace() with open('./data/hypervolumn.pickle', 'wb') as handle: pickle.dump(hv, handle)