Fishwick P.A. (editor) Handbook of Dynamic System Modeling

Подождите немного. Документ загружается.

9-12 Handbook of Dynamic System Modeling

TransferEntityFedSoAToFedDestC(P/S)

Parameter table

Interaction Parameter Datatype

Available

dimensions

Transportation Order

TransferEntityFedSoAToFedDestB Entity EntityType NA HLAreliable

TimeStamp

Fixed record datatype table

FieldRecord name

Name Type Semantics

Encoding Semantics

EntityName HLAASCIIString Name of the entity

Source HLAASCIIString FedSo

Destination HLAASCIIString FedDest

EntityType

Attributes Appropriate datatypes

(4,12)

The various attributes

of the entity

HLAfixedRecord An entity

Attributes

...

... ...

...

Appropriate datatypes

(4,12)

The various attributes

of the entity

Interaction class table

HLAinteraction Root(N) TransferEntity(N/S) TransferEntityToFedDest(N/S)

TransferEntityFedSoAToFedDestB(P/S)

FIGURE 9.5 OMT tables used for entity transfer speciﬁcation.

•

N (Neither): The federate is incapable of either publishing or subscribing to the interaction class.

HLAInteractionRoot is always this.

•

In an SOM, transferEntity will be

•

S if a federate wishes to get all entity transfer interactions.

•

N if the federate is not interested in receiving this global information.

•

In an SOM, transferEntityToFedDest will be

•

S if the federate’s model is FedDest.

•

N if the federate’s model is not FedDest (i.e., it is required to support the interaction class hierarchy

for a publish-only transferEntityFedSoToFedDest).

•

In an SOM, transferEntityFedSoToFedDest will be

•

P if the federate is FedSo, i.e., its model sends entities to FedDest’s model.

•

S if the federate is FedDest, i.e., its model receives entities from FedSo.

•

For an FOM, these interactions will be

•

transferEntity will be N if there is no “monitor” federate or S.

•

transferEntityToFedDest will be S.

•

transferEntityFedSoToFedDest will always be PS.

•

Classes designated as Subscribe or Neither are never sent, but they can have subclasses that are sent.

•

It will be assumed that when an FOM is composed from SOMs there will be some kind of entity

name resolution.

•

HLAinteractionRoot is a superclass of all other interaction classes in an FOM or SOM and is

mandatory.

The above is enough to deﬁne the representation of an entity transferred from one model to another via

the RTI. The translation of the datatype of this representation and the internal type representation of the

CSP must be performed by the CH according to the requirements of the CSP.

The interaction classes are meant to be used in the following way in a Type I IF. During initialization,

a federate will

•

indicate that it is capable of sending entities to various destination federates by publishing all

transferEntityFedSoToFedDest interactions, and

Distributed Modeling 9-13

•

indicate that it is capable of receiving entities from any other federate by subscribing to all

transferEntityToFedDest interactions.

During runtime, when the CSP sends the message equivalent to

output(entity, time, source, destination)

the CH will use destination to select the appropriate interaction class to use. It will then parameterize an

interaction instance with the details supplied in the output message. When the RTI passes an interaction

instance to the CH, the CH will use the instance’s details to pass the entity to the CSP in some input

message with source to indicate which model the entity has arrived from.

We now present an illustrative case study showing the use of the ETS to support the integration of

heterogeneous CSPs according to the Type I IRMs within the Type I IF.

9.6 Case Study

Consider the illustrative Type I IRM-based distributed simulation in Figure 9.6. A company manufactures

bicycles. Three models exist in three possibly heterogeneous CSPs that represent the three main parts of

the manufacturing system: a wheel production line (WPL), a frame production line (FPL), and a bicycle

assembly line (BAL) that assembles two wheels to one frame to produce a bicycle. The BAL checks wheels

for faults and can return them to the WPL for re-machining (an example of valid feedback for Type I

IRMs). Frames have no such problems. To describe part of the simulation, raw materials for the WPL

arrive every 20 min at entry point En1a and wait for processing in Q1a. When workstation W1a becomes

free, raw materials are taken from queue Q1a, processed into wheels and released. This activity takes a

ﬁxed time of 20 min. The newly created wheels then take 100 min travel time to be transferred to the BAL’s

entry point En3a. We assume that the entry point, the queue, and the workstation are adjacent. Frame

entities have the attributes “frame_size”which is of type integer and “frame_color” which is of type string.

Wheels have a single attribute “wheel_size” of type integer. The rest of the distributed simulation can be

described in a similar manner with the various times to perform actions shown on the models. Note that

in our example all distributions are ﬁxed. In a real simulation it is likely that these will be probabilistic

distributions and that there will be a greater number of model elements. However, our simple model

with its ﬁxed distributions is appropriate for purposes of illustration as we are concerned with distributed

simulation issues. Each simulation model in our example runs in a CSP or in different CSPs, with each

CSP/model combination a federation in our approach (the models WPL, FPL, and BAL and their CSPs

becoming federates Fd1, Fd2, and Fd3).

Figures 9.7(a)–9.7(c) show the SOMs for Fd1, Fd2, and Fd3. Figure 9.7(d) shows the composite FOM

for the federation as a whole. As can be seen, these tables provide a neutral representation of data that the

various heterogeneous CSPs are required to translate to and from as they send and receive entities. This

illustrates our contribution to emerging standards in this area in support of the Type I and II IRMs and

their IFs.

9.6.1 Illustrative Protocol

As part of the IF, the CH provides an interface consisting of a set of functions to be invoked by the CSP

when needed. Through the interface, the CH invokes necessary calls to the RTI ambassador on behalf of

the CSP and transfers the information received from the federate ambassador to the CSP. Figure 9.8 shows

the basic communication protocol between the CSP, CH, and RTI and its relationship with the ETS output

and input.

There are various different approaches to time management using a HLA RTI to support dis-

tributed discrete-event simulation (Fujimoto, 1998). The approach described here is based around

NextEventRequest (others are currently under investigation as part of the work developing the Type I IF).

9-14 Handbook of Dynamic System Modeling

Wheel production line (WPL)

fixed(20) fixed(20)

En1a Ex1aQ1a W1a

100

Raw materials

fixed(10)

En1b Ex1bQ1b W1b

100

Recycled wheels

Bicycle assembly line (BAL)

Wheels

En3a

Ex3b

Rejected

wheels

Inspection

fixed(5)

Q3a

Q3bW3a

75%

25%

Ex3aW3b

100

Frames

En3b

Q3c

Frame production line (FPL)

fixed(40) fixed(10)

En2 Ex2Q2 W2

100

Raw materials

fixed(20)

assembly Bicycles

FIGURE 9.6 The bicycle factory model.

Interaction class table (4.3)

HLAinteractionRoot(N) TransferWheelEntity(N) TransferWheelEntityToBAL(N) TransferWheelEntityWPLToBAL(P)

TransferWheelEntityToWPL(S)

TransferWheelEntityBALToWPL(S)

Parameter table (4.5)

Interaction Parameter Datatype

Available dimensions

Transportation Order

TransferWheelEntityBALToWPL

WheelEntity WheelType NA HLAreliable TimeStamp

TransferWheelEntityWPLToBAL

WheelEntity WheelType NA HLAreliable TimeStamp

Fixed record datatype table (4.12)

FieldRecord name

Name Type Semantics

Encoding Semantics

Wheel HLAASCIIString Name of the entity

Source HLAASCIIString WPL or BAL

Destination HLAASCIIString WPL or BAL

WheelType

Wheel_Size HLAinteger16BE Size of the wheel

HLAfixedRecord An Entity

Interaction class table

(a)

(b)

HLAinteractionRoot(N) TransferFrameEntity(N) TransferFrameEntityToBAL(N) TransferFrameEntityFPLToBAL(P)

Parameter Table

Interaction Parameter Datatype Available dimensions Transportation Order

TransferFrameEntityFPLToBAL FrameEntity FrameType NA HLAreliable TimeStamp

Fixed record datatype table

Field

Name

Record name

Type Semantics

Encoding Semantics

Frame HLAASCIIString Name of the entity

Source HLAASCIIString FPL

Destination HLAASCIIString BAL

Frame_Size HLAinteger16BE Size of the frame

FrameType

Frame_Color HLAASCIIString Color of the frame

HLAfixedRecord An Entity

FIGURE 9.7 Entity transfer speciﬁcations. (a) Wheel production line SOM; (b) frame production line SOM.

Distributed Modeling 9-15

TransferFrameEntity(N)

TransferWheelEntityToWPL(N) TransferWheelEntityBALToWPL(P)

TransferFrameEntityToBAL(S)

TransferWheelEntityFPLToBAL(S)

Parameter table

Interaction Parameter Datatype

Available dimensions

Transportation Order

TransferEntityBALToWPL

WheelEntity WheelType NA HLAreliable TimeStamp

TransferEntityFPLToBAL

FrameEntity FrameType NA HLAreliable TimeStamp

TransferEntityWPLToBAL

WheelEntity WheelType NA HLAreliable TimeStamp

Fixed record datatype table

FieldRecord name

Name Type Semantics

Encoding Semantics

Frame HLAASCIIString Name of the entity

Source HLAASCIIString FPL

Destination HLAASCIIString BAL

FrameType

WheelType

Frame_Size HLAinteger16BE Size of the frame

HLAfixedRecord An entity

Frame_Color HLAASCIIString Color of the frame

Wheel HLAASCIIString Name of the entity

Source HLAASCIIString WPL or BAL

Destination HLAASCIIString WPL or BAL

Size HLAinteger16BE Size of the wheel

(c)

Interaction class table

HLAinteractionRoot(N) TransferWheelEntity(N) TransferWheelEntityToBAL(S) TransferWheelEntityWPLToBAL(S)

TransferFrameEntity(N)

TransferWheelEntityToWPL(S) TransferWheelEntityBALToWPL(PS)

TransferFrameEntityToBAL(S)

TransferFrameEntityFPLToBAL(PS)

Parameter table

Interaction Parameter Datatype

Available dimensions

Transportation Order

TransferWheelEntityBALToWPL

WheelEntity WheelType NA HLAreliable TimeStamp

TransferFrameEntityFPLToBAL

FrameEntity FrameType NA HLAreliable TimeStamp

TransferWheelEntityWPLToBAL

WheelEntity WheelType NA HLAreliable TimeStamp

Fixed record datatype table

FieldRecord name

Name Type Semantics

Encoding Semantics

Frame HLAASCIIString Name of the entity

Source HLAASCIIString FPL

Destination HLAASCIIString BAL

FrameType

WheelType

Frame_Size HLAinteger16BE Size of the frame

HLAfixedRecord An entity

Frame_Color HLAASCIIString Color of the frame

Wheel HLAASCIIString Name of the entity

Source HLAASCIIString WPL or BAL

Destination HLAASCIIString WPL or BAL

Size HLAinteger16BE Size of the wheel

(d)

Interaction class table

HLAinteractionRoot(N) TransferWheelEntity(N) TransferWheelEntityToBAL(S) TransferWheelEntityWPLToBAL(PS)

FIGURE 9.7 Entity transfer speciﬁcations. (c) Bicycle assembly line SOM; (d) Bicycle manufacturing, system FOM.

When the CSP wishes to advance to the time of its next event, it issues an advanceTime request to the

CH. The CH invokes the corresponding RTI service nextEventRequest. The response from the RTI is zero

or many ETS interactions received via receiveInteraction and a new simulation time granted via timeAd-

vanceGrant. The interactions represent the arrival of entities at the time granted by timeAdvanceGrant and

may be less than the time initially requested by the CSP (i.e., entities arrive before the time of the original

next event—the new time of next event is that of the arriving entities). If no interactions appear, the time

granted is exactly the requested time. Either way, this grant time is returned to the CSP with the entities

received (if any) via input(entity, time, source), the CSP advances its local simulation time and continues

execution. If, as a consequence of this, any entities leave the simulation model, the CSP will send to CH as

9-16 Handbook of Dynamic System Modeling

COTS simulation package COTS handler Runtime infrastructure

advanceTime

newSimulationTime

nextEventRequest

receiveInteraction*

timeAdvanceGrant

sendinteraction*

Input(...)*

Output(...)*

FIGURE 9.8 Illustrative protocol.

many output(entity, time, source, destination) as appropriate. CH will translate these into ETS interactions

and then forward these to the RTI by invoking sendInteraction. This continues until some terminating

condition is met.

9.7 Conclusion

The purpose of this chapter has been to introduce some of the contemporary innovations in the use of

distributed computing techniques to support modeling. Our discussion has been centered around CSP-

based distributed simulation and represents a rich area of research that continues to be very rewarding for

those involved. As presented in the introduction, the development of a generalized solution to this area

will open up many new opportunities for those involved in dynamic systems modeling. Indeed several

successes have been achieved already. Brailsford et al. (to appear) describe work on the modeling of large

systems in health care with a distributed version of the CSP Simul8. Lendermann et al. (to appear) describe

progress in using a distributed version of the CSP AutoSched to model the rapidly changing demands of

semiconductor manufacture.

Acknowledgments

Sections of this chapter are based on the paper: Taylor, S.J.E., Wang, X., Turner, S.J. and Low, M.Y.H. 2006.

Integrating heterogeneous distributed COTS discrete-event simulation packages: An emerging standards-

based approach. IEEE Transactions on Systems, Man and Cybernetics: Part A, 36, 1, 109–122. The author

would also like to acknowledge the members of the CSPI-PDG for their hard work in developing the

current range of standards offered by the group.

References

Balci, O., 1988. The implementation of four conceptual frameworks for simulation modeling in high-level

languages. In Proc. 1998 Winter Simulation Conference, pp. 287–295.

Banks, J. (Ed.), 1998. Handbook of Simulation: Principles, Methodology, Advances, Applications, and Practice,

New York, NY: Wiley.

Distributed Modeling 9-17

Bapat, V. and D.T. Sturrock, 2003. The arena product family: Enterprise modeling solutions. In Proc. 2003

Winter Simulation Conference, pp. 210–217.

Bell, P.C., 1985. Visual interactive modeling in operational research: Successes and opportunities, Journal

of the Operational Research Society, vol. 36, pp. 975–982.

Bell, P.C. and R.M. O’Keefe, 1987. Visual interactive simulation—History, recent developments, and major

issues, Simulation, vol. 49, pp. 109–116.

Boer, C.A., A. Verbraeck,and H.P.M. Veeke, 2002a. Distributed simulation of complex systems: Application

in container handling. In Proc. 2002 EUROSIW, 02E-SIW-03.

Boer, C.A., A.Verbraeck, and H.P.M. Veeke, 2002b. The possible role of a backbone architecture in real-time

control and emulation. In Proc. 2002 Winter Simulation Conference, pp. 1675–1682.

Bosilj-Vuksic,V., M.I. Stemberger, J. Jarklic, and A. Kovacic, 2003. Assessment of E-business transformation

using simulation modeling, Simulation, vol. 78, pp. 731–744.

Brailsford, S., K. Katsaliaki, N. Mustafee, and S.J.E. Taylor, Accepted. Modelling very large complex systems

using distributed simulation: A pilot study in a healthcare setting. In Proc. Simulation Workshop

2006.

Bruzzone, A.G., 2003. Preface to modeling and simulation methodologies for logistics and manufacturing

optimization, Simulation, vol. 80, pp. 119–120.

Buxton, J.N. and J.G. Laski, 1962. Control and simulation language, The Computer Journal, vol. 5,

pp. 194–199.

Demirci, E., 2003. Simulation modelling and analysis of a port investment, Simulation, vol. 79, pp. 94–105.

Eldabi, T., R.J. Paul, and S.J.E. Taylor, 2000. Simulating economic factors in adjuvant breast cancer

treatment, Journal of the Operational Research Society, vol. 51, no. 4, pp. 465–475.

Fujii, S., T. Kaihara, and H. Morita, 2000. A distributed virtual factory in agile manufacturing environment,

International Journal of Production Research, vol. 38, no. 17, pp. 4113–4128.

Fujimoto, R.M., 1998. Time management in the high level architecture, Simulation, vol. 71, no. 6,

pp. 388–400.

Fujimoto, R.M., 2000. Parallel and Distributed Simulation Systems,NewYork,NY:Wiley.

Fujimoto, R.M., 2003. Distributed simulation systems. In Proc. 2003 Winter Simulation Conference,

pp. 124–134.

Gan, B.P., L. Li, S. Jain, S.J. Turner, C. Wentong, and J.H. Wen, 2000. Distributed supply chain simulation

across enterprise boundaries. In Proc. 2000 Winter Simulation Conference, pp. 1245–1251.

Gan, B.P. and S.J. Turner, 2000. An asynchronous protocol for virtual factorysimulation on shared memory

multiprocessor systems, Journal of Operational Research Society, vol. 51, pp. 413–422.

Goel, S., D.R. Strong, N. Richards, and N.C. Goel, 2002. A simulation-based method for the process to

allow continuous tracking of quality, cost, and time, Simulation, vol. 78, pp. 330–337.

Harrell, C.R. and R.N. Price, 2003. Simulation modeling using ProModel technology. In Proc. 2003 Winter

Simulation Conference, pp. 175–181.

Hibino, H., Y. Fukuda, Y. Yura, K. Mitsuyuki, and K. Kaneda, 2002. Manufacturing adapter of distributed

simulation systems using HLA. In Proc. 2002 Winter Simulation Conference, pp. 1099–1107.

Hofmann, M.A., 2004. Criteria for decomposing systems into components in modeling and simulation:

Lessons learned with military simulations, Simulation, vol. 80, pp. 357–365.

Hurrion, R.D., 1998. Visual interactive meta-simulation using neural networks, International Transactions

in Operational Research, vol. 5, pp. 261–270.

IEEE 1278, 1995. Standard for Distributed Interactive Simulation (DIS),NewYork, NY: Institute of Electrical

and Electronics Engineers.

IEEE 1516, 2000. IEEE Standard for Modeling and Simulation (M&S) High Level Architecture (HLA),New

York, NY: Institute of Electrical and Electronics Engineers.

IEEE 1516.0, 2000. IEEE Standard for Modeling and Simulation (M&S) High Level Architecture (HLA)—

Rules, New York, NY: Institute of Electrical and Electronics Engineers.

IEEE 1516.1, 2000. IEEE Standard for Modeling and Simulation (M&S) High Level Architecture (HLA)—

Federate Interface Speciﬁcation

, New York, NY: Institute of Electrical and Electronics Engineers.

9-18 Handbook of Dynamic System Modeling

IEEE 1516.2, 2000. IEEE Standard for Modeling and Simulation (M&S) High Level Architecture (HLA)—

Object Model Template (OMT) Speciﬁcation, New York, NY: Institute of Electrical and Electronics

Engineers.

IEEE 1516.3, 2004. IEEE Standard for Modeling and Simulation (M&S) High Level Architecture (HLA)—

Federate Development Process (FEDEP), New York, NY: Institute of Electrical and Electronics

Engineers.

Krahl, D., 2003. Extend: An interactive simulation tool. In Proc. 2003 Winter Simulation Conference,

pp. 188–196.

Lane, D.C., 1999. Social theory and system dynamics practice, European Journal of Operational Research,

vol. 113, pp. 501–527.

Law, A.M. andW.D.Kelton, 2000. Simulation Modeling and Analysis (3rd ed.), NewYork, NY: McGraw-Hill.

Lee, J.K., Y.H. Lim, and S.D. Chi, 2004. Hierarchical modeling and simulation environment for intelligent

transportation systems, Simulation, vol. 80, pp. 61–76.

Lendermann, P., B.P. Gan, and L.F. McGinnis, 2001. Distributed simulation with incorporated APS

procedures for high-ﬁdelity supply chain optimization. In Proc. 2001 Winter Simulation Conference,

pp. 1138–1145.

Linn, R.J., C.S. Chen, and J.A. Lozan, 2002. Development of distributed simulation model for the

transporter entity in a supply chain Process. In Proc. 2002 Winter Simulation Conference,pp.

1319–1326.

Low, Y.H., B.P. Gan, J.J. Wei, X. Wang, S.J. Turner, and W. Cai, 2001. Implementation issues for

shared state in HLA-based distributed simulation. In Proc. 2001 European Simulation Symposium,

pp. 5–13.

Mertins, K. and M. Rabe, 2002. Inter-enterprise planning of manufacturing systems applying simulation

with IPR protection. In Proc. 5th International Conference on Design of Information Systems for

Manufacturing (DIISM), pp. 149–156.

Mertins, K., M. Rabe, and F.W. Jäkel, 2000. Neutral template libraries for efﬁcient distributed simulation

within a manufacturing system engineering platform. In Proc. 2000 Winter Simulation Conference,

pp. 1549–1557.

McLean, C. and F. Riddick, 2000. The IMS MISSION architecture for distributed manufacturing

simulation. In Proc. 2000 Winter Simulation Conference, pp. 1539–1548.

Nance, R.E., 1996. A history of discrete event simulation programming languages, in History of

Programming Languages—II, New York, NY: Association for Computing Machinery Press, pp.

369–427.

Nance, R.E. and R. Sargent, 2002. Perspectives on the evolution of simulation, Operations Research, vol. 50,

no. 1, pp. 161–172.

Paul, R.J. and S.J.E. Taylor, 2002. What use is model reuse: Is there a crook at the end of the rainbow? In

Proc. 2002 Winter Simulation Conference, pp. 648–652.

Peterson, J.L., 1981. Petri Net Theory and the Modeling of Systems, New Jersey,: Prentice-Hall.

Pidd, M., 1998. Computer Simulation in Management Science (4th ed.), Chichester, UK: Wiley.

Rabe, M. and F.W. Jäkel, 2001. Non military use of HLA within distributed manufacturing scenarios. In

Proc. Simulation und Visualisierung, pp. 141–150.

Rabe, M. and F.W. Jäkel, 2003. On standardization requirements for distributed simulation, in Production

and Logistics. Building the Knowledge Economy, Twente, The Netherlands: IOS Press, pp. 399–406.

Robinson, S., 2002. Modes of simulation practice: Approaches to business and military simulation,

Simulation Practice and Theory, vol. 10, pp. 513–523.

Robinson, S., 2004. Simulation: The Practice of Model Development and Use, Chichester, UK: Wiley.

Robinson, S. 2005. Distributed simulation and simulation practice, Simulation, vol. 81, no. 1, pp. 5–13.

Robinson, S., R.E. Nance, R.J. Paul, M. Pidd, and S.J.E. Taylor, 2004. Simulation model reuse: Deﬁnitions,

beneﬁts and obstacles, Simulation Modeling Practice and Theory, vol. 12, pp. 479–494.

Rohrer, M.W., 2003. Maximizing simulation ROI with AutoMod. In Proc. 2003 Winter Simulation

Conference, pp. 201–209.

Distributed Modeling 9-19

Schriber, T.J. and D.T. Brunner, 2003. Inside discrete-event simulation software: How it works and why it

matters. In Proc. 2003 Winter Simulation Conference, pp. 113–123.

Straßburger, S., 2001. Distributed Simulation Based on the High Level Architecture in Civilian Application

Domains, Ghent, Belgium: Society for Computer Simulation International.

Straßburger, S., G. Schmidgall, and S. Haasis, 2003. Distributed manufacturing simulation as an enabling

technology for the digital factory, Journal of Advanced Manufacturing Systems, vol. 2, no. 1, pp. 111–

126.

Sudra, R., S.J.E. Taylor, and T. Janahan, 2000. Distributed supply chain management in GRIDS. In Proc.

2000 Winter Simulation Conference, pp. 356–361.

Taylor, S.J.E., 2003. HLA-CSPIF: The high level architecture—COTS simulation package interoperation

forum. In Proc. 2003 Fall Simulation Interoperability Workshop, 03F-SIW-126.

Taylor, S.J.E., B.P. Gan, S. Strassburger, and A. Verbraeck, 2003. HLA-CSPIF technical panel on distributed

simulation. In Proc. 2003 Winter Simulation Conference, pp. 881–887.

Taylor, S.J.E. and N. Mustafee, 2003. An analysis of internal/external event ordering strategies for COTS

distributed simulation. In Proc. 2003 European Simulation Symposium, pp. 193–198.

Taylor, S.J.E., S. Robinson, and J. Ladbrook, 2005. An investigation into the use of net-conferencing

groupware in simulation modeling, Journal of Computing and Information Technology, vol. 13,

no. 1, pp. 1–10.

Taylor, S.J.E., R. Sudra, T. Janahan, G. Tan, and J. Ladbrook, 2002. GRIDS-SCS: An Infrastructure for

Distributed Supply Chain Simulation, Simulation, vol. 78, pp. 312–320.

Tocher, K.D., 1963. The Art of Simulation, London, UK: English Universities Press.

Zülch, G., U. Jonsson, and J. Fischer, 2002. Hierarchical simulation of complex production systems by

coupling models, International Journal of Production Economics, vol. 77, pp. 39–51.

Model Execution

∗

Kalyan S. Perumalla

Oak Ridge National Laboratory

10.1 Introduction ................................................................ 10-1

Systems and Models

•

Elements of Execution

•

Execution Platforms

•

Generating Executables from

Models

•

Executable Timelines

•

Pacing the Execution

10.2 Time-Stepped Execution ............................................ 10-5

Example: Heat Equation

•

Parallelizing Time-Stepped

Execution

10.3 Discrete-Event Execution ........................................... 10-7

Execution Method

•

Example: ATM Multiplexer

•

Parallelizing Discrete-Event Execution

10.4 Summary ..................................................................... 10-13

10.1 Introduction

A computer software-based model is typically designed to produce a trace of system evolution over time.

The actual process of computing the model state and producing the state values as the simulation time

is advanced is called model execution. Models could be designed with a speciﬁc execution technique in

mind, or could be generally amenable to multiple execution techniques. Two popular methods that are

used to execute models are time-stepped method and discrete-event method. Each of these methods could

in turn be executed either sequentially (on a single processor), or in parallel (using multiple processors

concurrently). In this chapter, we describe the time-stepped and discrete-event execution methods and

outline some of the common approaches to their sequential and parallel execution. Execution concepts

common to the methods are described followed by implementation details of the methods.

10.1.1 Systems and Models

Figure 10.1 shows the generalarchitecture in which model execution is used. A system of interest is modeled

at appropriate resolution and the resulting models are conﬁgured for given scenarios of interest. The

execution is based on the given conﬁguration of interest and its execution generates the required output.

This chapter deals with different methods by which simulation models can be mapped to executable units

and how their execution is controlled to obtain meaningful and useful output from the simulation. Model

execution is multifaceted, involving issues such as execution order and synchronization.

∗

This manuscript has been authored by UT-Battelle, LLC, under contract DE-AC05-00OR22725 with the U.S.

Department of Energy. The UnitedStates Government retains and the publisher, by accepting the article for publication,

acknowledges that the United States Government retains a nonexclusive, paid-up, irrevocable, worldwide license to

publish or reproduce the published form of this manuscript, or allow others to do so, for United States Government

purposes.

10-1