Firstly, the standard OpenERP stocktake logic is a bit simple and has some limitations if you have serialised products.

Putting that to one side, you are faced with two choices.

1. Standard OpenERP does not have the concept of stock by bin location which is normal for large warehouses. If you want this then there is a v6 community module which is pretty good. My client that I was investigating this for decided stock by bin was too complex, so I haven't looked at in the last couple of years. If you install this module you will need to test the standare stock-take logic, and you will probably have to fix.

2. You go standard - this means that despite products being in multiple locations in your warehouse you treat them as one location (say stock) for on hand and count purposes. A quick and dirty half-way house would be to record the actual locations against the product (so product_id, location_id, bin_name) and print this onto your documentation.

Firstly, I'm new to OpenERP, and even ERP. I imagine if you have the same product stored in multiple locations within the warehouse, then there might be something wrong with the warehouse procedure itself. Otherwise, then I imagine you could create a module "warehouse locations", then remove the fields for product location, then add field to link to multiple locations (many2one, one2many) to the product.