(2) When the number of Kanbans is much greater than the number of workstations, a uniform allocation is nearly optimal.
Andijani and Clark [1] investigated a system similar to the one given by Mitra and Mitrani[14]. They considered the average throughput rate and the average work-in-process (WIP) as their steady state performance measures. Their objective was to allocate Kanbans in order to maximize the marginal bene®t attributed to throughput rate and WIP levels.