Hi Vishwanathan,
Here is below attached pdf for the derivation u can refer...
Actually what is given in the first pdf (see my first reply) the Inversion coefficient (IC)is for the EKV model of the MOS i.e. this model is valid in all regions of operaion of the MOS.
IC= Id/(Io (W/L)) for Io see the image attached its different for different technologies !
if
IC<0.1 then the MOS is in subthreshold (weak inversion) region.
0.1<IC<10 then it is in moderate inversion
IC>10 then the MOS is in strong inversion.
What u want to know is the MOS region of operation ...Id current u want in the MOS (its usually in uA when in strong inversion and nA when in weak inversion) ...and take the min value of L ..(for eg 180nm take L=0.18um )
Now u can determine the W of the transistors....U can also change the value of the L later and adjust to be able to accomodate all transistors....
This method is for EKV model of the transistors......if you are using the BSIM model then also you can apply this method to get the W/L of the transistors approximately....and can make intelligent conclusions.
Hope this helps !