The classical batch processing system ignores the cost of increased waiting time for users. Consider a single batch characterized by the following parameters:
M average mounting time
T average service time per job
N number of jobs
S unit price of service time
W unit price of waiting time per user
Show that the optimal batch size that minimizes the cost of service time and waiting time per user (within a single batch) is
Simply create a formula for the cost of service time and a formula for the cost of the waiting time per user. You'll find that these will be S/TN and WN/M respectively. Adding these gives you total cost. This is how that formula is created.