In my paper "different approach to design....." the formula is given to calculate the duty cycle of the boost converter........

The worst case is for a duty ratio of 0.5 so it would be a conservative choice to start with.

Thanks dear Christophe, but i think that i have to choose the value when the power is maximum. 

Do you mean input or output power ?

Indeed, the optimal duty ratio will vary if you intend to implement a MPPT strategy. This will depend on the lighting of the panel in that case. Thus the current ripple might be less than the case duty ratio = 0.5, but considering the data you gave, there is much unknown, hence my proposal as a conservative (and probably sub optimal) choice.

In MPPT, the input and output power must be almost equal

Actually, as you don't have any storage in you converter (apart from the filters, but they are negligible), it is also true even if you don't have MPPT. By input power, I meant the lighting condition, which in some cases may not reach the nominal values of the PV (in france at least, maybe you are more lucky in morroco !)

Well, as you are using boost converter Vo = (1/1-D)Vin when Vo is the converter DC output and Vin is the PV voltage and For sure D is the duty cycle.  so D= 1- (Vpv / Vo) and you design for specific voltage and current output range, and for an input voltage range which is for your pv panel will be within range 0 - Voc , and the switching frequency of your converter. So by applying these given parameters in the duty cycle formula above you can determine the range of D and in turns the maximum and minimum value of it and then you can get the right value for L , C and electrical C/C's of the planned switch as follow :

Dmin = 1 - (Vimax/Vomin)

Dmax = 1 - (Vimin/Vomax)

L > D * Vin * (1-D) / (freq * 2 * Iout )

peak inductor current ==> Ipk = (Vinmax * D)/(f * L)

Cap > Iout / (Vripple * freq)

Hope this helps you and you can find this in my publications in a practical experiment .

Yasser

Thank you for your answer