Rerun cerebro with same data: how to refresh Broker?

[Abdul Mokhammed]

Hello!
I need to optimize strategy parameters but for some reasons i can't use optstrategy

The problem i got is that trying to run cerebro with same data i still have position remaining open from previos run which leads to incoherency

The question is, how can i refresh all Broker attribute values back to initial before running again?

INIT_VAL = 1e06
strats = []
i = 1
l = len(combs_filtred)

cerebro = bt.Cerebro(cheat_on_open = False)
data = PandasData_feed(dataname = df)
cerebro.adddata(data, name = 'main')
cerebro.broker.setcommission(0.04/100)

for f, s, llqt, lw, slqt, sw, lp, sp, tp, sl, bbp, bbm in combs_filtred:
    print('{}\{} '.format(i,l), end = '')
    cerebro.broker.setcash(INIT_VAL)
    strat = cerebro.addstrategy(
        LiqStrategy,
        log = False,
        opt_mode = True,
        fast=f,  # period for the fast moving average
        slow=s,   # period for the slow moving average
        long_lq_thresh = llqt,
        long_window = lw,
        short_lq_thresh = slqt,
        short_window = sw,
        long_period = lp,
        short_period = sp,
        tp_size = tp,
        sl_size = sl,
        bb_period = bbp,
        bb_mult = bbm)
    
    t = dt.datetime(2022, 1, 20, 22, 15)
    cerebro.run()
    strats.append(strat)
    i += 1

[noufbouf]

Create a new instance of cerebro ? basicaly run this line again : cerebro = bt.Cerebro(cheat_on_open = False)

[bartjs]

@abdul-mokhammed

I would suggest getting optstrategy to work, rather than re-running new instances of cerebro, try to have optstrategy run, and post your code and errors here.