How? Such that the voltage ripple on the input side stays within 'acceptable limits'.

Without knowing what the supply looks like, and also not knowing which ripple would be considered 'acceptable' in your application: impossible to give a value or a reasonable formula.

Depends as well on switching frequency, the inductance of the choke, the output power, and so on ...

Dear Dreher, thank you !

Oh, if you need some kind of 'world formula': I doubt such a thing is available. Subsets are available, wur a 'world formula' including several types of supplies (single phase, three-phase, DC with characteristics) ? ? ?

Maybe you should try to develop it ?

Can't find what ? The generic 'world formula' ? I doubt it is available.

You are trying what exactly ?

If, perchance, you are interested in improving available boost converter circuits, then you are wasting your time, on the input capacitor.  The dominant problem is with an element, not shown on your schematic, the core loss in the inductor. this can be modeled as a shunt resistor, but a resistor is linear, and the core loss is anything but.

Tremendous improvement is immediately available by new ferrite shapes, wide multiple gaps, with coherent winding over gaps and all. None of this is well described by simple algebra, but because the boost circuit is are so widely used for power factor correction, the  potential energy saving is important to the health of the planet.

I use simple rule: value of C1 more or equal C2*V2/V1. Fine value need other parameters evaluate.

Saad, C1 is not required if impedance of power source is very low but really the capacitor, close connected to leads of switch and L1, have impedance to lower. Additionally in general case we have not zero length of wires. I evaluate the impedance and the discharge capability of C1 for not optimal values of impedance of power source. Same electrical charge dq is transferred from C1 to C2 over L1 in one period. C1 ripple voltage (without power source) dV1=C1/dq. I estimate roughly it impact to C2 ripple voltage via coefficient V2/V1. See links

http://www.ti.com/lit/an/slta055/slta055.pdf

https://www.monolithicpower.com/Portals/0/Documents/Products/Documents/appnotes/AN051_MP2130.pdf

Thank you 

I liked that & thank you