Group replacement policies for machines operating in parallel are gene
rally evaluated by computing the associated expected cost per unit tin
e. However; as with all stochastically evolving systems, variance shou
ld play an important managerial role. Mo matter which policy is used b
y the maintenance manager, knowledge of the associated variance provid
es important information about the variability of the costs that will
be incurred by using this policy. Additionally, there will be cases wh
ere a decision maker is willing to incur increased expected cost per u
nit time in order to reduce the variability of the costs among cycles.
This paper demonstrates that it is feasible and practicable to calcul
ate the variance of the cost per unit time associated with group maint
enance policies.