multi input buck-boost system questions

[Sniper1]

So here is my scenario:
I have multiple Li ion cells in parallel, each with a buck boost (bidirectional)  + INA219 for current and voltage measurement  , each SMPS is tuned by a DAC so i have programmable 3-5V
all cells are separated by FETs , i also have a common rail for 5V in and another one for DC out.
I know how to charge the cells in CC CV mode but the discharge part is the problem for me.

How can i discharge more then 1 cell at .5A ( or wherever) into the output buss?
In my mind it seems doable since P=UI where U is set by the cell and I by me so P is limited in a way (this is for the input side of the SMPS)
As for the output it is again P=UI  and P= delta P , here i somehow need to make all SMPS output the desired Power (since i need to have a set U, probably 4.5-5v since I need it to be higher then cell voltage in order for the IC to allow reverse current flow and this is a feature and not a hack since TI itself made a app note about this, for super caps), since P is tied to the other side then the only thing i can adjust is the current aka my load.
Note: yea there are loses and perfect .5A is not that important 
And yes i am very careful with Li ion cells since i know that they are dangerous, that is why i am playing at low currents and only with 1 cell as of now 

Am i on the right track? Or am i overthinking stuff and there is a much simpler way ?

[mtwieg]

Some more background might be helpful. Sounds like a lot of complexity if you just want to store electrical power, as opposed to an off the shelf battery pack with cells in series. Perhaps your intent is to make it so you can swap cells in and out arbitrarily?

[Sniper1]

here is more context:
it is a 3 step process :
1 charge n cells in parallel from a 5V DC source (do it at CC CV and program that to be a code variable) charge ends at 4.2V or 3.6V ( also code select able for individual cells) more or less done , cell start at a random charge level and some will be bad
2 discharge them at .5A CC and measure the discharge curve ( and compute Ah) this is the step that is related to my post, i need to discharge m cells in parallel where m<=n , my output is a common DC bus and at the end of that buss i am also designing the load ( that will be another DC DC hopefuly).
3 repeat step 1, but now all cells are at about 3V