118
5 Development of a Simulation-based Methodology …
has to be determined very precisely which parameters can be created for the optimization process in the simulation model. Additionally, optimization constraints
are required to include given restrictions and limitations of the real production
system in the optimization process. The consideration of important constraints is
usually done using the simulation model to depict side conditions. Generally, it is
easier doing this depiction in the model rather than formulating systems of inequations. At this point, the example of the already mentioned minimum retention
time in a machine state is used again. The minimum retention time may be required to complete cooling processes or system shut downs before the machine can
be restarted again. In the simulation model, the minimum retention time can be
inserted easily into the state graph of the machine logic by dividing a state into a
fixed and a variable part. The minimum retention time thus becomes an element
of the machine logic that must be run through within the given state changes as
soon as the machine switches into this state. The formulation of equations for this
restriction is much more complicated. It is necessary to clarify which parameters
are included in the formulation of the restrictions, how they can be determined
exactly, which range of variation and which units of measure they have. Furthermore, it must be defined under which conditions the state may be left again,
etc. This makes the formulation of the restrictions in a mathematical model very
time-consuming and complex.
The definition of restrictions should always be carried out in parallel with
the definition of the optimization parameters, as the value range of the parameters is directly limited by the restrictions. The scheduling of orders for example
has a direct impact on the production output of a shift, start times of production
orders should therefore not be used as a parameter for increasing energy efficiency, unless there are clear restrictions defined, limiting the range of allowed
changes for the order rescheduling by a few minutes forwards/backwards to avoid
load peaks for example. The same applies to the use of the optimization parameter “machine state”. Soberly, the idle state is always less energy intensive than the
productive state, as well are the standby or even the off mode. Without the definition of restrictions and conditions under which machine state changes may take
place, the whole optimization process does not work. Without the clear specification in which order which machine state can be reached, which states have to
be passed through to get into the productive state and which minimum retention
times must be met in certain states during the non-value-adding machine times,
a minimization of the total energy of a production would always lead to putting
machines into the off state immediately after completing a production job.
A detailed example description for the definition of optimization parameters
and constraints will be given in the fictional case study as well as in the practical
5 Development of a Simulation-based Methodology …
has to be determined very precisely which parameters can be created for the optimization process in the simulation model. Additionally, optimization constraints
are required to include given restrictions and limitations of the real production
system in the optimization process. The consideration of important constraints is
usually done using the simulation model to depict side conditions. Generally, it is
easier doing this depiction in the model rather than formulating systems of inequations. At this point, the example of the already mentioned minimum retention
time in a machine state is used again. The minimum retention time may be required to complete cooling processes or system shut downs before the machine can
be restarted again. In the simulation model, the minimum retention time can be
inserted easily into the state graph of the machine logic by dividing a state into a
fixed and a variable part. The minimum retention time thus becomes an element
of the machine logic that must be run through within the given state changes as
soon as the machine switches into this state. The formulation of equations for this
restriction is much more complicated. It is necessary to clarify which parameters
are included in the formulation of the restrictions, how they can be determined
exactly, which range of variation and which units of measure they have. Furthermore, it must be defined under which conditions the state may be left again,
etc. This makes the formulation of the restrictions in a mathematical model very
time-consuming and complex.
The definition of restrictions should always be carried out in parallel with
the definition of the optimization parameters, as the value range of the parameters is directly limited by the restrictions. The scheduling of orders for example
has a direct impact on the production output of a shift, start times of production
orders should therefore not be used as a parameter for increasing energy efficiency, unless there are clear restrictions defined, limiting the range of allowed
changes for the order rescheduling by a few minutes forwards/backwards to avoid
load peaks for example. The same applies to the use of the optimization parameter “machine state”. Soberly, the idle state is always less energy intensive than the
productive state, as well are the standby or even the off mode. Without the definition of restrictions and conditions under which machine state changes may take
place, the whole optimization process does not work. Without the clear specification in which order which machine state can be reached, which states have to
be passed through to get into the productive state and which minimum retention
times must be met in certain states during the non-value-adding machine times,
a minimization of the total energy of a production would always lead to putting
machines into the off state immediately after completing a production job.
A detailed example description for the definition of optimization parameters
and constraints will be given in the fictional case study as well as in the practical