dike_model_simulation.py

_# -*- coding: utf-8 -*-
"""
Created on Wed Apr 24 11:18:24 2019

@author: ciullo
"""

from ema_workbench import (MultiprocessingEvaluator,
                           Scenario, SequentialEvaluator, Policy)

from ema_workbench.em_framework.samplers import sample_uncertainties, sample_levers
from ema_workbench.util import ema_logging
import pickle
import time
import pandas as pd
from problem_formulation_4b import get_model_for_problem_formulation


def build_Pol(df, pf):
    policies = []
    for i, row in df.iterrows():
        name = '{}_{}'.format(pf, str(i))
        decision = {lever: row[lever] for lever in df.columns}
        policies.append(Policy(name=name, **decision))
    return policies

if __name__ == '__main__':
    ema_logging.log_to_stderr(ema_logging.INFO)

    pf = 'disaggregated'

    dike_model = get_model_for_problem_formulation(pf)[0]
        
    average_values = {'Bmax': 150, 'Brate': 1.5, 'pfail': 0.5}
    scen1 = {}

    for key in dike_model.uncertainties:
        dikename, unc = key.name.split('_')
        scen1.update({key.name: average_values[unc]})

    ref_scenario = Scenario('reference', **scen1)
    policies = sample_levers(dike_model, 500)

    # No dike increase, none of the rfr project is selected, no damage
    # reduction:
    dike_increase = dict(zip((91, 101, 201, 301, 401, 501),
                             (0, 0, 0, 0, 0, 0)))

    names = {'DikeIncrease': 0, 'rfr': 0,
             '201.0': 0, '401.0': 0}
    ## zero policy:
    pol0 = {}

    for key in dike_model.levers:
        s1, s2 = key.name.split('_')
        if s2 == 'DikeIncrease':
            branch = s1.split('.')[0]
            pol0.update({key.name: dike_increase[int(branch)]})

        else:
            pol0.update({key.name: names[s2]})

    policy0 = Policy('Policy 0', **pol0)
    policies = [policy0]
# pf1
#    df = pd.read_excel('./results_pf2_100000nfe____.xlsx'
#                       )[[l.name for l in dike_model.levers]]
#    policies.extend(build_Pol(df, 'pf1'))

## pf2
#    df = pd.read_excel('./results_pf1_100000nfe_NSB1e6_0.xlsx'
#                       )[[l.name for l in dike_model.levers]]
#    policies.extend(build_Pol(df, 'pf2'))
#
#### pf3
#    df = pd.read_excel('./results_pf3_100000nfe.xlsx'
#                       )[[l.name for l in dike_model.levers]]
#    policies.extend(build_Pol(df, 'pf3'))



## SIMULATION:
## single run
#    start = time.time()
#    dike_model.run_model(ref_scenario, policies[1])
#    end = time.time() - start
#    print(end)
#    results = dike_model.outcomes_output

##     series run
#    with SequentialEvaluator(dike_model) as evaluator:
#        results = evaluator.perform_experiments(ref_scenario, policies[1])

## multiprocessing
#    with MultiprocessingEvaluator(dike_model, n_processes = 3) as evaluator:
#        results = evaluator.perform_experiments(ref_scenario, policies)
###
#### save results
#    with open("./results//FirstDecentResults/simulation_noncooppolsx100.p".format(pf), "wb") as f:
#            pickle.dump(results, f)