摘要:何謂系統模擬? 優點和缺點
(1) Definition of Simulation:
A controlled statistical sampling technique (experiment) that is used,
in conjunction with a model, to obtain approximate answers for questions
about complex, multifactor probabilistic problems. It is most
useful when analytical and numerical techniques cannot supply answers.
Simulation experiments study a probabilistic system by sampling the
random behavior of an unknown output distribution which is an implicit
function, represented by the computer code based on system
logics, of the random variates generated from the assumed input
distributions.
(2) Advantages of Simulation:
• Faithful adherence to the real system’s details—easy (compared to
analytical solution which often requires stringent assumptions)
• Simulation experiments permit inferences to be drawn about systems
(A) without building the systems, if they are only proposed systems
(B) without disturbing the systems, if they are operating systems
that are costly or unsafe to experiment with
(C) without destroying the systems, if the objective of an experiment
is to determine their limits of stress
(3) Disadvantages of Simulation:
• Need for specialized training
• Inability to optimize: simulation can only identify the probable best
alternative among those in the set considered
• Only generate estimate with sampling error (ˆ 6= )
• Costly computer runs (Var(ˆ)) =) statistics