into two intermediate demand matrices 
and 
: for the first leg (usually the auto part) of the trip, and for the second leg (usually the transit part). These intermediate matrices can be obtained as follows by using the before-mentioned substitutions in (8) and summing over p or q:
Fortunately, the explicit knowLEDge of the variable 
is also not required for carrying out the solution algorithm, it suffices to know the values for the total parking lot usage for each lot, which can easily be obtained from either first or second leg demand matrix as
The model and the solution algorithm described in the previous sections have been implemented in the EM
