Transcript Introduction to Discrete-Event Simulation COMP8700 Agent-Directed Simulation Dr. Levent Yilmaz

COMP8700 Agent-Directed Simulation
Introduction to Discrete-Event Simulation
Dr. Levent Yilmaz
M&SNet: Auburn M&S Laboratory
Computer Science & Software Engineering
Auburn University, Auburn, AL 36849
http://www.eng.auburn.edu/~yilmaz
© Yilmaz- - 2004-12-06
“Introduction to Discrete-Event Simulation”
1
Agent-Directed Simulation
Use of Simulation for Agents
Agent Simulation: simulation of agent systems.
Agent simulation is already used in several categories of
application areas such as:
Engineering applications: electrical engineering,
irrigation systems, manufacturing systems, mechatronics,
networks, robotics, software, as well as
transportation/logistics;
Management/economy applications: economy, ecommerce, and management;
Social systems and human behaviour applications:
social systems, psychology/human behaviour,
physiology, negotiation, and organization theory;
Environment applications: ecosystems, land use;
Military applications.
© Yilmaz- - 2004-12-06
Use of Agents for Simulation
Agent-supported Simulation: use of agents in a simulation study for at
least one of the following purposes: (in a computer-aided simulation
environment –including simulation-based problem solving
environments):
- to provide computer assistance for front-end and/or back-end
interface functions;
- to process elements of a simulation study symbolically (for example,
for consistency checks and built-in reliability); and
- to provide cognitive abilities to the elements of a simulation study,
such as learning or understanding abilities.
Agent-based Simulation: use of agents for the generation of model
behavior in a simulation study
“Introduction to Discrete-Event Simulation”
2
Aim
• The aim of this lecture is to overview the fundamentals,
underlying principles, conceptual frameworks, and the
life-cycle of a discrete-event simulation study.
© Yilmaz- - 2004-12-06
“Introduction to Discrete-Event Simulation”
3
© Yilmaz- - 2004-12-06
“Introduction to Discrete-Event Simulation”
4
© Yilmaz- - 2004-12-06
“Introduction to Discrete-Event Simulation”
5
© Yilmaz- - 2004-12-06
“Introduction to Discrete-Event Simulation”
6
© Yilmaz- - 2004-12-06
“Introduction to Discrete-Event Simulation”
7
© Yilmaz- - 2004-12-06
“Introduction to Discrete-Event Simulation”
8
© Yilmaz- - 2004-12-06
“Introduction to Discrete-Event Simulation”
9
© Yilmaz- - 2004-12-06
“Introduction to Discrete-Event Simulation”
10
© Yilmaz- - 2004-12-06
“Introduction to Discrete-Event Simulation”
11
© Yilmaz- - 2004-12-06
“Introduction to Discrete-Event Simulation”
12
© Yilmaz- - 2004-12-06
“Introduction to Discrete-Event Simulation”
13
© Yilmaz- - 2004-12-06
“Introduction to Discrete-Event Simulation”
14
© Yilmaz- - 2004-12-06
“Introduction to Discrete-Event Simulation”
15
© Yilmaz- - 2004-12-06
“Introduction to Discrete-Event Simulation”
16
© Yilmaz- - 2004-12-06
“Introduction to Discrete-Event Simulation”
17
© Yilmaz- - 2004-12-06
“Introduction to Discrete-Event Simulation”
18
© Yilmaz- - 2004-12-06
“Introduction to Discrete-Event Simulation”
19
© Yilmaz- - 2004-12-06
“Introduction to Discrete-Event Simulation”
20
© Yilmaz- - 2004-12-06
“Introduction to Discrete-Event Simulation”
21
© Yilmaz- - 2004-12-06
“Introduction to Discrete-Event Simulation”
22
© Yilmaz- - 2004-12-06
“Introduction to Discrete-Event Simulation”
23
© Yilmaz- - 2004-12-06
“Introduction to Discrete-Event Simulation”
24
© Yilmaz- - 2004-12-06
“Introduction to Discrete-Event Simulation”
25
© Yilmaz- - 2004-12-06
“Introduction to Discrete-Event Simulation”
26
© Yilmaz- - 2004-12-06
“Introduction to Discrete-Event Simulation”
27
• The initialization include the
assignment of initial values to all
the attributes.
• Fixed time increment is as the
time flow mechanism.
• If an activity’s condition is
satisfied, the actions of that
activity are executed.
© Yilmaz- - 2004-12-06
“Introduction to Discrete-Event Simulation”
28
© Yilmaz- - 2004-12-06
“Introduction to Discrete-Event Simulation”
29
• The three phase approach
combines activity scanning
and event scheduling
frameworks.
• Unlike the two-phased
approach the time flow
mechanism implements a
variable time increment
approach.
• All B (bound to occur –
unconditional) activities with
occurrence times equal to the
simulation clock are
executed before activity
scanning takes place.
© Yilmaz- - 2004-12-06
“Introduction to Discrete-Event Simulation”
30
• Under this framework the life
cycle of an object that moves
through and interacts with the
processes of the system is
represented.
• Initialization phase creates the
attributes and dynamic objects.
• In the clock update phase the
time is advanced to the move
time of the front-end object of
the FOL.
• All objects with equivalent
move time are transferred to
COL, which is processed in the
scan phase.
© Yilmaz- - 2004-12-06
“Introduction to Discrete-Event Simulation”
31
• Problem formulation is the process by
which the initially communicated
problem is translated into a formulated
problem sufficiently well defined to
enable specific research action.
• Not enough attention is given in the life
cycle.
• The accuracy of problem formulation
greatly affects the acceptability and
credibility of simulation results.
© Yilmaz- - 2004-12-06
“Introduction to Discrete-Event Simulation”
32
© Yilmaz- - 2004-12-06
“Introduction to Discrete-Event Simulation”
33