Why channel is multiplied with signal in BPSK over rayreigh BUT convolve in OFDM

In matlab simulation, why channel h is multiplied by transmitted signal Xt in BPSK over rayleigh --> h*Xt
but in "OFDM", channel will be convolved with transmitted signal --> conv(h,Ht)

why?

why we don't do convolve in BPSK over rayleigh...or may be just do normal multiply in OFDM?

I think but not sure that in OFDM the signal when it reaches the channel is in time domain (cause it went through IFFT) and the channel object is also in time doamin , that is why we have to do a convolution of teh two, but if you want to have easier code you can first FFT the channel and multiply it with theincoming signal when it is still in frequency domain
hope that helps

its a model.

For the BPSK model, you really should perform a convolution -- or actually apply the same channel effects to both cases. When the channel has no multipath (for wireless) and has ample bandwidth (for cabled communications), the channel largely becomes an attenuation of the signal and accounts for lower SNR. If the channel is frequency selective (eg, multipath or bandwidth limits), then the channel also results in echos or dispersive behavior of the signal, as seen in convolution.

For the BPSK system where such effects matter, an adapative equalizer would be common. OFDM and the variants (OFDMA/SCFDMA) become attractive because each sub-channel is in a very narrow bandwidth, over which the channel is approximately flat. The degree to which this occurs can only be seen if you include the frequency-selective channel in the first place.

is h a transfer function or impulse response ? Actually, if h is an impulse response you should convolve in BPSK and multiply in OFDM, in BPSK, the modulation symbols run in the time domain while in OFDM, modulation symbols are applied to IFFT and the BPSK then becomes a frequency domain sequence.

"permute" is right... its a model...but a bit of correction here: channel is convolved in both the cases of BPSK and OFDM!!
Fading comes from multipath phenomena. i.e multiple copies of the same signal arriving from different paths. constructive/destructive addition on the reciever leads to amplitude fluctuations which we call fading. now answer to the question.
if all the multipath components arrive at almost the same time, then the channel is modelled as a single tap channel(or filter). if viewed as a signal, h will look like an impulse function of some amplitude. Convolution of impulse function is same as multiplication... i.e as done in BPSK(called flate fading)..[not forgetting the bandwidth factor explained by "permute"]