@Markiyan Hirnyk OK thanks.

@Markiyan Hirnyk Very much thanks. But what if one would like to plot beyond beta = 1e4 say 1e6 or higher. Oscillations are still there. Is there a way to deal with taht? Thanks.

F

@Preben Alsholm Thanks a lot for your clarifications. I get exactly that is I was loking for! I will go through your code to get more insights about what you did.

Best greetings,

F

@Preben Alsholm Very much thanks for your powerful answer. The approach you suggest seems to be perfect for this problem. What do yo mean by "and sampled the result at Chebyshev zeros, after which the integration was performed and solved for the coefficients..."? Could you please post your answer with some more details so that I can view what you mean? Thanks.

@Preben Alsholm OK then. s takes values between 0 and 1.

@Axel Thank you for answering the question. Yes I get your point. The integral is therefore divergent.

Bests

@acer Dear Acer,

thank you very much for your powerful answer. Now the computations are much faster than before. Furthermore, It tured out that the Fourier transforms can drastically be simplified. Bests

a

@tomleslie Thank you very much for your answer. So it worked for you without problem. I use actually Maple 18 for these computations and the memory completely saturates for N = 6 or more ! Now I desided to integrate up to 40 (which is almost infinity because of exponential.) It might be a version problem as you just pointed out.

Thanks again for your time and effort.

a