However, knowledge of the function cR (t) is a precondition
for optimization. The presented model assumes that the course
of this function will be based on the results of observation of a
group of vehicles in service (several hundreds of vehicles).
The whole time of subsystem service will be divided into a
final number of the same time periods and in each of these
periods, the cumulative costs expended for repairs of a given
subsystem in all vehicles will be observed.
From these costs, the average cost per vehicle in each period
of service can be easily established, and from them, a number
of discrete values representing the time development of
cumulative costs for repair of the subsystem can be
determined. These discrete values can be, for the purpose of
further solution, approximated by a suitable function CR (t)
(e.g. by a method of least square). This function can be used to
obtain the average unit repair cost cR (t) and the
instantaneous unit cost for repairs uR (t) . Fig. 4 shows an
example of using discrete values to develop the cumulative
costs for repairs function.