Hi,When using MonteCarlo to price continuous barrier options one needs to do a correction due to the fact that simulaiton is discrete whereas the monitoring is continuous.There are several ways to do this. One is the Broadie-Glasserman-Kou(BGK) barrier "shift" approach.My question is if anyone has any expereice of using this together with Local Volatiltiy Model.I mean should one a) stick to a static BS volatility, e.g. from the barrier value, which is simple and stable b)or use the local volatilities for the correction which can compliate things and introduce noiseAlso is using local volatilities in the BGK correction even theoretically justifiable? Thanks in advance.K

During MC's, I would simply apply a Brownian bridge correction using the local vol sig(S,t).

Yeah, I would just use the "left" vol as the constant vol.