A GPSS World Model is a sequence of Model Statements. A Model Statement may be a GPSS Statement or a PLUS Procedure definition. A GPSS Statement is either a Command or a Block Statement.
PLUS is an acronym for the Programming Language Under Simulation, GPSS World’s embedded programming language. PLUS Experiments are a kind of Procedure.
This chapter describes the use of Model Statements in the GPSS World simulation environment. It then discusses the general structure of GPSS Statements. Individual GPSS Commands are discussed in detail in Chapter 6, and Block Statements are in Chapter 7. Of the PLUS Statements, only PROCEDURE and EXPERIMENT are Model Statements. PLUS is described in detail in Chapter 8.
Before you can run a simulation, you must Create a Simulation Object. This is called the Initial Model Translation. It’s easy. You just open a Model Window on the Model Object, and
CHOOSECommand / Create Simulation
You will be told if any corrections need to be made, due to syntax errors.
When you have Translated a model successfully, the Simulation Object automatically gets a copy of the Translated Model, and executes any Commands embedded in the Model Files. You can now open any of the online windows upon the simulation. If you have put one or more START Commands into one of the Model Files, the Simulation Object executes them in order with the other Commands, without waiting to be told to.
Any Model Statement can be part of the Initial Model Translation, or can be sent to an existing simulation later, as an Interactive Statement.
When a Model is Translated successfully, its Block Structure is created, a Procedure Registration List is created, a Command List is created, and the whole thing is sent to the Simulation Object.
When a model is Translated, the Block sequence is preserved, and separately, the initial Command sequence is preserved. When the Simulation Object receives a Command it either performs it immediately, or places it on the Command Queue to be dealt with in turn.
After the Initial Model Translation, you may send any Model Statement to the existing Simulation Object. These statements are called Interactive Statements. They are themselves Translated, and sent to the Simulation Object, which receives and executes them.
What happens depends on what kind of Model Statement is sent. PROCEDURE and EXPERIMENT Statements define or redefine a PLUS Procedure in the simulation. Commands are either executed immediately (SHOW or HALT), or are placed on the simulation’s Command Queue, to be dealt with in sequence.
Interactive Block Statements are executed immediately in Manual Simulation Mode. This means that a temporary Block is created and the Active Transaction attempts to enter it. Then, the Block is destroyed.
The Block structure of the simulation is set when the Model Object is Translated. It is not altered until the next Retranslation. During the Model Translation, as the Translator encounters each Block Statement, it appends a representation of it to the simulation’s Block structure.
After a simulation is STARTed, it is the original sequence of Block Statements that determines the sequence of actions in the simulation. Unless the flow of Transactions is modified, each Active Transaction attempts to enter one Block after the other in the original sequence.
Block Statements received by the simulation after the original Model File was Translated, are NOT incorporated into the Block structure of the simulation. Instead, they cause a temporary block to be created, used, and destroyed in Manual Simulation Mode.
Commands are either Immediate or Queued. Only HALT and SHOW are Immediate; all others are Queued. When the Simulation Object receives an Immediate Command it executes it right away. Queued Commands, on the other hand are placed in the Simulation’s Command Queue. The Simulation Object executes one Queued Command after the other, occasionally interrupted by Immediate Commands.
A Command List is created by the Initial Model Translation and sent to the Simulation Object, just as Interactive Commands are sent. As the Translator encounters Commands during the Initial Model Translation, they are placed on the list. When the whole model has been Translated, the Command List is sent to the Simulation Object after the Block structure and the Procedure Registration List.
START, CONTINUE, and CONDUCT are Queued Commands. When the Simulation Object takes one off the Command Queue, the simulation runs until its Termination Count is nonpositive. More Commands may be waiting to be done after that.
The CONDUCT Command is used to run Experiments, which are a special kind of PLUS Procedure. Experiments are registered in a Simulation Object like any other PLUS Procedure, and they can have arguments passed to them by the CONDUCT Command.
The HALT Command is a special case. Not only is it an Immediate Command, but it has the effect of clearing off all Commands from the Command Queue. When you send a HALT Command to the Simulation Object, the simulation is brought to the Halted State, and all activity on behalf of the simulation comes to an end. You are then free to control the simulation as you please.
During the Initial Model Translation, when the Translator encounters a Procedure or Experiment Statement, it puts the Translated Procedure on the Procedure Registration List. The Simulation Object registers all the user-defined Procedures when it receives the Translated model.
When the Simulation Object receives an Interactive Procedure Statement, it incorporates the Translated Procedure into the simulation’s Procedure Library, replacing any existing Procedure of the same name.
The invocation of a Procedure occurs when an Expression with a Procedure Call, or a Procedure Call Statement, in a PLUS Procedure is executed. Invocations can be performed interactively, as well, by placing a Procedure invocation in the operand of a SHOW Command.
Experiments can only be invoked by CONDUCT Commands. Procedures invoked during an Experiment invocation have a special property: they may call the DoCommand library procedure. This is discussed in Chapter 8.
All GPSS World Objects can be saved and opened in a later session. Models, Simulations, Reports, and Text Objects can all be modified and saved at any time. The File / Open Menu Command is available to bring the object back.
A Simulation can be saved at any point in its run for the purposes of debugging, demonstration, replay, or simply to save an intermediate state of the Simulation Object
The act of saving a Simulation or Experiment causes a HALT Command to be sent to the Simulation Object. In the case of an Experiment, you will need to save the state of the completed part of the Experiment in Global Variables and Global Matrices if you don't want to restart the Experiment from the beginning when you reopen it.
There are two kinds of GPSS Statements: Block Statements, which cause a GPSS Block Entity to be created, and Commands, which do not. Commands either define a non-Block Entity or cause an action to occur. PLUS Statements are not discussed here.
Unlike a PLUS Statement, a GPSS Statement must reside on a single line of text, consisting of up to 250 characters.
GPSS Statements are composed of parts called fields. A field is a variable number of printable characters terminated by white space or a delimiter. Although the exact composition of a Statement varies a little, in general, a Statement is laid out as follows: Line number (optional) | Label (optional) | Verb (required) | Operands (depend on verb) | Comment (optional).
Arbitrary Line numbers are retained for compatibility purposes. If used, they must begin in column 1. However, they are ignored by the Translator. A line number given in an error message is an absolute cardinal number, denoting the place of the line in the Model File, and is not a user-selected line number.
The commands are described in detail in Chapter 6. They are:
BVARIABLE - Define a Boolean Variable Entity.
CLEAR - Reset statistics and remove Transaction.
CONDUCT - Execute a registered PLUS Experiment.
CONTINUE - Resume the simulation.
EQU - Assign a value to a User Variable.
EXIT - End the GPSS World Session.
FUNCTION - Define a Function Entity.
FVARIABLE - Define an Fvariable Entity.
HALT - Stop the simulation and delete all Queued Commands.
INCLUDE - Read and Translate a secondary Model File.
INITIAL - Initialize or modify a Logicswitch, Savevalue, or Matrix Entity.
INTEGRATE - Automatically integrate a time differential in a User Variable.
MATRIX - Define a Matrix Entity.
QTABLE - Define a Qtable Entity.
REPORT - Set the name of the Report File or request an immediate report.
RESET - Reset the statistics of the simulation.
RMULT - Set the seeds of the first 7 Random Number Generators
SHOW - Evaluate and display Expression.
START - Set the Termination Count and begin a simulation.
STEP - Attempt a limited number of Block entries.
STOP - Set a stop condition based on Block entry attempts.
STORAGE - Define a Storage Entity.
TABLE - Define a Table Entity.
VARIABLE - Define a Variable Entity.
A Block Statement creates a GPSS Block. When a Transaction enters a Block Entity, a special action occurs which is determined by the nature of the Block. Block Statements may be inserted into a Model Object using the fullscreen text editor or using a Block Creation Dialog. It is also possible to use the text editor to place Block Statements into Text Objects that can be referred to by an INCLUDE Statement in the Text Object. The Block structure of the simulation is set when the Model Object is Translated. After a simulation is STARTed, it is the original sequence of Block Statements that determines the sequence of actions in the simulation. Unless the flow of Transactions is modified, each Active Transaction attempts to enter one Block after the other in the original sequence. Block Statements received by the simulation after the original Model File was Translated, are NOT incorporated into the Block structure of the simulation. Instead, they cause a temporary block to be created, used, and destroyed in Manual Simulation Mode, which is discussed above.
The GPSS Block Statements are described, in detail, in Chapter 7. They are:
ADOPT - Change Assembly Set.
ADVANCE - Place Transaction on Future Events Chain.
ALTER - Test and modify Transactions in a Group.
ASSEMBLE - Wait for and destroy related Transactions.
ASSIGN - Modify Transaction Parameter.
BUFFER - Place Transaction on the Current Events Chain behind its priority peers.
CLOSE - End the Data Stream.
COUNT - Place count of entities into a Transaction Parameter.
DEPART - Decrement content of a Queue Entity.
DISPLACE - Change the Next Sequential Block of a Transaction.
ENTER - Occupy or wait for storage units in a Storage Entity.
EXAMINE - Test group membership.
EXECUTE - Perform action specified by a different Block.
FAVAIL - Change status of a Facility Entity to "available".
FUNAVAIL - Change status of a Facility Entity to "not available".
GATE - Test entity and modify Transaction flow.
GATHER - Wait for related Transactions.
GENERATE - Create Transaction and place on Future Events Chain.
INDEX - Modify Transaction Parameter.
INTEGRATION - Turn the integration of a User Variable On or Off.
JOIN - Place a member into a Numeric or Transaction Group.
LEAVE - Release storage units of a Storage Entity.
LINK - Move Transaction to Userchain Entity.
LOGIC - Modify Logicswitch Entity.
LOOP - Decrement Parameter, jump to different Block if result is non zero.
MARK - Place value of system clock into Transaction Parameter.
MATCH - Wait for related Transaction to reach conjugate MATCH Block.
MSAVEVALUE - Assign value to Matrix Entity element.
OPEN - Initialize a Data Stream.
PLUS - Evaluate PLUS Expression and save result in Parameter.
PREEMPT - Displace Facility owner.
PRIORITY - Modify Transaction priority.
QUEUE - Increment content of a Queue Entity.
READ - Bring the next line of data from a Data Stream.
RELEASE - Free Facility Entity.
REMOVE - Take a member out of Numeric or Transaction Group.
RETURN - Free Facility Entity.
SAVAIL - Change status of Storage Entity to "available".
SAVEVALUE - Assign a value to Savevalue Entity.
SCAN - Test Transaction group, place value in Parameter.
SEEK - Change the line pointer in a Data Stream.
SEIZE - Assume ownership of or wait for a Facility Entity.
SELECT - Place selected entity number into Transaction Parameter.
SPLIT - Create related Transaction.
SUNAVAIL - Change status of Storage Entity to "not available".
TABULATE - Update Table Entity.
TERMINATE - Destroy Transaction, decrement Termination Count.
TEST - Test arithmetic condition and modify Transaction flow.
TRACE - Set Trace Indicator of the Active Transaction.
TRANSFER - Move to specified Block.
UNLINK - Remove Transaction from Userchain Entity.
UNTRACE - Turn off Trace Indicator in the Active Transaction.
WRITE - Send a value to a Data Stream.
The parts of a GPSS Statement are called fields. GPSS Statements are of free form in the sense that you do not have to worry about lining up fields with column numbers, except for line numbers. The format is much the same for all statements, with some fields being optional for some verbs. In general, the fields of a statement are laid out left-to-right in the following sequence:
·Line number (optional). Ignored.
·Label (variable, depends on verb field).
·Operands, including any conditional operator (variable, depends on verb).
Line numbers are ignored by GPSS World. They are permitted only in GPSS Statements, and if used, they must begin in column 1. They are retained for compatibility purposes.
The label field allows you to name and refer to entities with names of your choosing. A label must obey the rules of name construction described later in this chapter. You may assign your own values to names with EQU Commands and PLUS Assignment Statements. This is useful when you want to experiment with several values of some variable. However, it is easiest to let GPSS World use its own unique internal numbers when you name entities and Transaction Parameters. If you choose to assign your own numbers to entity names, you should assign the value in an EQU statement before you define or refer to the entity by name. If you assign your own values to names, then you are responsible for avoiding reference conflicts.
When GPSS World encounters a new name, it assigns a unique integer to the name. These integers are distinct and contiguous starting at 10,000. You may use the same name to refer to different entity types. For example You may have a Storage Entity and a Table Entity both named Motorpool. However if you use the name Motorpool as a label in a second STORAGE Command, the old Storage Entity will be replaced with a new one. Permanent Block Entities, however, are not replaced in this manner.
Blocks are a special case. An interactive Block Statement is executed as a temporary Block in Manual Simulation Mode. If you need to be able to change the characteristics of a Block in the middle of a simulation, you can redefine the named values used in operands using EQU Statements. You can also use an EXECUTE Block and modify its operand interactively. Blocks cannot be inserted interactively. A change to the permanent Block structure requires a retranslation.
You may not use a number as a name in a Label Field. That would violate naming conventions. If you want to explicitly control the number of an entity, you can assign a value to a name using an EQU statement before using the name as a label.
You may not use an SNA as a Block Label.
The Verb Field follows the Label Field and requires a GPSS keyword. It must be one of the GPSS Block names or one of the Commands. A list of valid Verbs is available through the online help facility.
Most Verbs are followed by one or more operator and operand fields. These fields depend on the verb. Some are required and some are optional. The Online Help facility is always available to show you which syntax is acceptable. When you want to finish or skip an operand, you should press a delimiter key. Traditionally, operands are separated by commas. If all required information in a statement has been provided, you may skip optional fields.
The forms you may use in any given Operand Field are listed with the specific Statement descriptions in Chapters 6 and 7. This information is available as online help, as well.
There are two kinds of comment fields. First, by starting a statement with ; or *, the whole statement is considered to be a comment. Second, an optional comment field follows the operand field of any statement. Function followers and statements with expression fields do not have comment fields.
When all required operands of a statement have been completed, you can press; to begin a comment. At this time you can enter a comment of both upper and lower case ASCII characters.
Expressions in GPSS World are defined as part of the PLUS Language. You can use them in GPSS Statement Operands and in PLUS Procedures.
Some Block operands permit the use of parenthesized PLUS Expressions. The list of acceptable syntactic variables will then include ParenthesizedExpression as one or the operand’s alternative forms. Some Commands do not need to use the outer parentheses, but if you always parenthesize PLUS Expressions used in GPSS Statements, you will be safe.
It’s easy to go between PLUS and GPSS. Most GPSS Statement Operands allow you to use a PLUS Expression. This means that you can provide a simple factor, a small calculation, or a PLUS Procedure invocation right there in the Operand. On the other hand, by using Expressions containing SNAs, you can access GPSS entity state variables, as well as global User Variables, from within PLUS Procedures.
Section 8.3.2 of this manual tells you how to build PLUS Expressions. All the building blocks except System Numeric Attributes are discussed there. We discuss SNAs in the next section.
System Numerical Attributes, or SNAs, are simulation "state variables" that are available for use through a simulation. They return numeric or string values, and may be used in GPSS Statement operands and in Expressions.
Some SNAs are maintained automatically, and others require a calculation when they are called. For example, the accumulated count of entries into a Block is available as an SNA. There are over 50 SNA classes, each of which will contribute to the power of your simulations. It is wise to become familiar with all of them. Many SNAs, such as PR, are evaluated on behalf of the active Transaction. This means that some attribute of a Transaction is necessary in order to evaluate the SNA. Such SNAs cannot be evaluated unless there is an Active Transaction. For example, you cannot refer to Transaction Parameters in the operands of GENERATE Blocks because when a Transaction enters a GENERATE Block for the first time, it has no Parameters. Other SNAs, such as the system clock C1, can be evaluated without referring to a Transaction.
SNA Entity Specifiers
Most SNAs may be specified in one of many forms, beginning with the SNA class. For example the SNA W22 returns the number of Transactions waiting in Block number 22. The Block identifier in this case is the number 22, but there are several other possibilities. This SNA could have been specified using any of the following entity specifiers:
·Wj - where j is a positive integer, the number of the Block in the simulation.
·W$Name - where Name is the location of the desired Block.
·W*j - where j is a positive integer, the number of the Parameter of the active Transaction which contains the number of the desired Block. This is indirect addressing.
·W*Name - where Name is the name of the Parameter of the Active Transaction which contains the number of the desired Block. This is indirect addressing.
·W*$Name - where Name is the name of the Parameter of the Active Transaction which contains the number of the desired Block. The $ is not needed and is used only as a separator. Essentially, this is identical to W*Name. This is indirect addressing.
·W*Parameter - denotes that either the W*j, the W*Name, or the W*$Name may be used.
The Matrix Entity SNA class MX is a special case. It may contain up to 3 indirect addresses. For example the SNA MX*Sales(*Partnumber,*January) locates the Matrix Entity whose number is in the Transaction parameter named Sales, and then retrieves the element whose row and column numbers are in the Parameters named Partnumber and January, respectively. Normally, the Active Transaction would have already passed through three ASSIGN blocks initializing the Parameters Sales, Partnumber, and January with the numbers or name values desired.
Some SNA classes, such as A1, AC1, C1, M1, MP, PR, and TG1, are "atomic" SNAs. Atomic SNAs are complete and do not need an entity name or number to complete the evaluation of the SNA
When you are in any operand field, the valid SNA forms are available in an online help message.
The following SNAs are available for use in the operand fields and Expressions of commands and statements. In all cases Entnum must be replaced by any entity specifier. The entity specifier could be a name (preceded by a$ separator) or number, or for indirect addressing, it could be an asterisk,*, followed by a name or number. The formal rules for building operands are given in the Appendix.
Here is a complete list of SNAs available in GPSS World:
·A1 - Assembly Set of the Active Transaction. Integer value.
·AC1 - Value of absolute system clock. Simulated time since last CLEAR. Real value.
·BVEntnum - Result of evaluating Boolean Variable Entity Entnum. Real value.
·C1 - Value of relative system clock. Simulated time since last RESET. Real value.
·CAEntnum - Average Userchain content. The time weighted average number of chained Transactions for Userchain Entnum. Real value.
·CCEntnum - Total Userchain entries. The count of all Transactions chained to Userchain Entnum. Integer value.
·CHEntnum - Current Userchain content. The current number of Transactions chained to Userchain Entnum. Integer value.
·CMEntnum - Maximum Userchain content. The maximum number of Transactions chained to Userchain Entnum. The "high water mark". Integer value.
·CTEntnum - Average Userchain residence time. The average duration of Transactions at Userchain Entnum. Real value.
·FEntnum - Facility busy. If Facility Entity Entnum is currently busy, FEntnum returns 1. Otherwise FEntnum returns 0. Integer value.
·FCEntnum - Facility capture count. The number of times Facility Entity Entnum has been SEIZEd or PREEMPTed by a Transaction. Integer value.
·FIEntnum - Facility Entnum interrupted. If Facility Entity Entnum is currently preempted, FIEntnum returns 1. Otherwise FIEntnum returns 0. Integer value.
·FNEntnum - Function. Result of evaluating Function Entity Entnum . Real value.
·FREntnum - Facility utilization. The fraction of time Facility Entity Entnum has been busy. FREntnum is expressed in parts-per-thousand and therefore returns a value 0-1000, inclusive. May be nonintegral. Real value.
·FTEntnum - Average Facility holding time. The average time Facility Entity Entnum is owned by a capturing Transaction. Real value.
·FVEntnum - Facility in available state. FVEntnum returns 1 if Facility Entity Entnum is in the available state, 0 otherwise. Integer value.
·GNEntum - Numeric Group count. GNEntnum returns the membership count of Numeric Group Entnum. Integer value.
·GTEntnum - Transaction Group count. GTEntnum returns the membership count of Transaction Group Entnum. Integer value.
·LSEntnum - Logicswitch set. LSEntnum returns 1 if Logicswitch Entity is in the "set" state, 0 otherwise. Integer value.
·MBEntnum - Match at Block. MBEntnum returns a 1 if there is a Transaction at Block Entnum which is in the same Assembly Set as the Active Transaction. MBEntnum returns a 0 otherwise. Integer value.
·MPParameter - Transit Time, Parameter. Current absolute system clock value minus value in Parameter Parameter. Real value.
·MXEntnum(m,n) - Matrix Savevalue. The value in row m, column n of Matrix Entity Entnum is returned. In Matrices of more than 2 dimensions, all other indices are assumed to be 1. Unlike MX class SNAs, PLUS Expressions can refer to any element of a higher dimension matrix.
·M1 - Transit time. M1 returns the absolute system clock minus the "Mark Time" of the Transaction. Real value.
·NEntnum - Block entry count. The total number of Transactions which have entered Block Entnum is returned. Integer value.
·PParameter or *Parameter - Parameter value. PParameter or *Parameter returns the value of Parameter Parameter of the Active Transaction. Integer, real, or string value.
Indirect addressing uses the notation SNA*Parameter.
·PR - Transaction priority. The value of the priority of the Active Transaction. Integer value.
·QEntnum - Current Queue content. The current count value of Queue Entity Entnum. Integer value.
·QAEntnum - Average Queue content. The time weighted average count for Queue Entity Entnum. Real value.
·QCEntnum - Total Queue entries. The sum of all entry counts for Queue Entity Entnum. Integer value.
·QMEntnum - Maximum Queue content. The maximum count of Queue Entity Entnum. The "high water mark". Integer value.
·QTEntnum - Average Queue residence time. The time weighted average of the count for Queue Entity Entnum. Real value.
·QXEntnum - Average Queue residence time excluding zero entries. The time weighted average of the count for Queue Entity Entnum not counting entries with a zero residence time. Real value.
·QZEntnum - Queue zero entry count. The number of entries of Queue Entity Entnum with a zero residence time. Integer value.
·REntnum - Available storage capacity. The storage content (or "tokens") available for use by entering Transactions at Storage Entity Entnum. Integer value.
·RNEntnum - Random number. RNEntnum returns a random integer 0-999 from Random Number Generator Entnum. Integer value.
·SEntnum - Storage in use. SEntnum returns the amount of storage content (or "tokens") currently in use by entering Transactions at Storage Entity Entnum. Integer value.
·SAEntnum - Average storage in use. SAEntnum returns the time weighted average of storage capacity (or "tokens") in use at Storage Entity Entnum. Real value.
·SCEntnum - Storage use count. Total number of storage units that have been acquired from Storage Entity Entnum. Integer value.
·SEEntnum - Storage empty. SEEntnum returns 1 if Storage Entity Entnum is completely available, 0 otherwise. Integer value.
·SFEntnum - Storage full. SFEntnum returns 1 if Storage Entity Entnum is completely used, 0 otherwise. Integer value.
·SREntnum - Storage utilization. The fraction of total usage represented by the average storage in use at Storage Entity Entnum. SREntnum is expressed in parts-per-thousand and therefore returns a value 0-1000, inclusive. May be nonintegral. Real value.
·SMEntnum - Maximum storage in use at Storage Entity Entnum. The "high water mark". Integer value.
·STEntnum - Average holding time per unit at Storage Entity Entnum. Real value.
·SVEntnum - Storage in available state. SVEntnum returns 1 if Storage Entity Entnum is in the available state, 0 otherwise. Integer value.
·TBEntnum - Nonweighted average of entries in Table Entity Entnum. Real value.
·TCEntnum - Count of nonweighted table entries in Table Entity Entnum. Integer value.
·TDEntnum - Standard deviation of nonweighted table entries in Table Entity Entnum. Real value.
·TG1 - Remaining Termination Count. TG1 returns the count that is decremented by TERMINATE Blocks with a positive A operand. This value is initialized by START Statements and indicates completion of the simulation when it is 0. Integer value.
·VEntnum - Result of evaluating arithmetic or floating point Variable Entity Entnum. Real value.
·WEntnum - Current Block count. The current number of Transactions in Block Entity Entnum. Integer value.
·XEntnum - Savevalue. The value of Savevalue Entity Entnum is returned. Integer, real, or string value.
·XN1 - Active Transaction number. The Transaction number of the Active Transaction is returned. Integer value.
·Z1 - Free memory. Value returned by the Operating System. Integer value.
Special rules apply when a simulation is run in GPSS/PC Compatibility Mode. This are discussed more fully below.
Operators are used to combine data elements in Expressions. Data are coerced if a specific type is required by the operator. For example, if a numeric operation is to be performed on a variable with a string value, the numeric equivalent derived from the characters in the string is used.
A Setting is available in the Simulation Page of the Settings Notebook which switches the roles of the [*] operator and the [#] operator.
The operators used in GPSS World are:
^Exponentiation. A^B returns A raised to the power of B.
# (or *)Multiplication. A # B returns the numeric product of A and B.
/Division. A / B returns the quotient of A divided by B.
\Integer Division. A \ B returns the result of integer division of A by B.
@Integer Remainder. A @ B returns the modulo division of A by B.
-Subtraction. A - B returns the difference of A and B.
+Addition. A + B return the sum of A and B.
>=‘GE’ Greater Than or Equal To. A >= B returns 1 if A is numerically greater than or equal to B, 0 otherwise.
<=‘LE’ Less Than or Equal To. A <= B returns 1 if A is numerically less than or equal to B, 0 otherwise.
>‘G’ Greater Than. A > B returns 1 if A is numerically greater than B, 0 otherwise.
<‘L’ Less Than. A < B returns 1 if A is numerically less than B, 0 otherwise.
=‘E’ Equal A = B returns 1 if A is numerically equal to B, 0 otherwise.
!=‘NE’ Not Equal. A != B returns 1 if A is numerically different from B, 0 otherwise.
& ‘AND’ Logical And. A & B returns 1 if and only if both A and B are non zero, 0 otherwise.
| ‘OR’ Logical Or. A ‘OR’ B returns 1 if A or B or both are non zero, 0 otherwise.
When Expressions are evaluated, the precedence of operators determines the order of operations. An operator with a higher precedence is evaluated before an operator with a lower precedence. The precedence of operators is as follows, from highest to lowest:
# (or *) / \Multiplication, Division, Integer Division
- +Addition, Subtraction
>= <= > <Comparison Operators
= !=Equal, Not Equal
All entities have positive entity numbers. When you refer to an entity by name, the value of the name is matched with the entity number of the entity. Indirect addressing means that the entity number is in a Transaction parameter. To use indirect addressing, you must have an asterisk, [*] (or possibly [#] -- see Settings), in the operand. For example, the SNA Q*2 uses the value of parameter 2 as the Queue Entity number. This SNA returns the current content of that Queue Entity.
You must move an entity number (or name value) into a Transaction parameter before the SNA is evaluated.
Indirect addressing uses the notation SNA*Parameter. As another example, to SEIZE the Facility whose number is the value returned by the function whose number is in a parameter named Tellerselector, you would type SEIZE FN*Tellerselector .
In GPSS World, intermediate results and SNAs are never truncated unless you explicitly do so. However, for compatibility purposes, intermediate result truncations are retained in GPSS/PC Compatibility Mode. When simulations are run this way, SNAs always return truncated integers (unless used as Function Modifiers), and the intermediate results obtained while evaluating Bvariable Entities and Variable Entities, are truncated.
The use of GPSS/PC Compatibility Mode is discussed in Chapter 1.
In Compatibility Mode, when an Expression is used to define an SNA, its result is always truncated to an integer. You should choose the time units of the simulation so that these truncations are not significant.
In Compatibility Mode, the evaluation of an Expression is affected by the Command it is used in.
·SHOW performs all operations as double precision floating point numbers and display the result in the data window as a double precision floating point number.
·VARIABLE creates a GPSS arithmetic variable which, when evaluated by an SNA call, performs all operations as double precision floating point numbers, then truncates all intermediate results before proceeding to the next operation. Finally it truncates and returns the overall result.
·FVARIABLE creates a GPSS "floating point" variable which, when evaluated by an SNA call, performs all operations as double precision floating point numbers, then truncates the result. Fractional values may be considered by an appropriate choice of units, such as "thousandths". Then an SNA value of 500 is interpreted as the number 1/2.
·BVARIABLE creates a GPSS Boolean variable which, when evaluated by an SNA call, performs all operations as double precision floating point numbers, then returns 1 if the result is non zero, 0 otherwise.
· FUNCTION Commands, of types other than C, evaluate the argument and select a list member based on the result to determine the final value, which is truncated.
· C Type FUNCTION evaluation begins with the evaluation of the argument. The result is always an integer, and it is used to identify the line segment of the Function. The argument is then used in a double precision linear interpolation to arrive at the double precision result of the Function. If the argument is an RN class SNA, a random fraction between 0 and .999999, inclusively, is used. If the Function is not used as a "Function Modifier" the final result is truncated.
· SNAs, except continuous Functions used as Function Modifiers, return an integer.
You can create names to refer to Blocks, User Variables, or other GPSS Entities. Names must begin with an alphabetic character, and may contain up to 200 alphabetic and numeric characters, and underscores [ _ ].
Your primary tasks in selecting a name are, first, to avoid using a Keyword, System Numeric Attribute, or System Numeric Attribute Class, and second, to use something meaningful that can be remembered.
A name may not be a verb or partial verb, a keyword, or a valid SNA. If you are unsure of which names are illegal, you will be safe if you include an underscore [ _ ] in the names you create. Names beginning with at least 3 letters and then a digit are also safe. In any case, GPSS World will not let you create an invalid name.
Names used in the Label field of a GPSS Statement are used to refer to GPSS Entities, and are called Entity Labels, or just Labels. Block Labels are also called Locations.
Locations are assigned values corresponding to the Block Entity number. You should not use a Block location name for any other purpose. GPSS World will cause an Error Stop to occur if you attempt to do so. This eliminates the possibility of inadvertently altering which Block is associated with a label.
When a Simulation Object receives a name from a Session, it assigns a unique numeric value to the name. These system-defined numbers start at 10,000. You can assign your own value, of any data type, later.
If you want to force a named entity to have an entity number of your choosing, you must use the name in an EQU Statement before you define (or reference) the entity by name. It is then up to you to avoid different names inadvertently referring to the same entity. You may use a single name to refer to one entity of each entity type. For example, the SNAs Q$BARBER and F$BARBER refer to different entities. The first refers to a Queue Entity, the second to a Facility Entity. There is no confusion when the same name refers to different entity types.
When an entity is created with a label, it takes the current value of the label to keep as its permanent Entity Number. Even if the Named Value is later changed, the Entity Number of the created entity is not.
If it can, the Simulation Object will create a GPSS entity when it first encounters a name reference to that entity type. However, entities that cannot be created without additional information, such as Storage Entities, must be defined by GPSS Commands before they are referenced. The entities which must be defined before they are referenced are listed in Chapter 4.
In addition to using Names as Labels, you can use them to represent your own values. Such variables are called User Variables, and are created by their occurrence in an EQU Command or in a PLUS Assignment Statement. They serve the traditional purpose of programming variables. A user created name which is first assigned a value with an EQU Command, but later appears as an Entity Label, is still known as a Label.
User Variables can be used to hold numeric or string values during a simulation. They can be referenced in PLUS Expressions, and can be altered by an EQU Command or an assignment Statement in a PLUS Procedure. In addition, they can be automatically updated by integration. This is discussed in Chapter 4, under the section Continuous Simulation.
Except for temporary names, all user-created names are known throughout the whole model. The exceptions are those names appearing in the TEMPORARY or the TEMPORARY MATRIX Statements of a PLUS Procedure. New instances of temporary Named Values or Matrices are created when the PLUS Procedure is invoked and deleted when the PLUS Procedure exits. These names override any other like-named items in the Model during execution of the PLUS Procedure. Other names used in PLUS Procedures refer to entities known globally throughout the model.
Within an Experiment, you should place the results of simulations into Global Variables and Global Matrices in order to preserve them across simulations. For example, the ANOVA library procedure requires a Global Matrix Entity to be passed as an argument. Also, the DoCommand library procedure is available during an Experiment. Any strings passed to DoCommand should not include the names of TEMPORARY and TEMPORARY MATRIX variables or of procedure arguments, because the string is Translated in a Global Scope where temporary variables are not recognized.
Numbers may be stored internally in any of three data types, integer, real, and string. Most variables can take on any of these three data types. Since numeric operators coerce their operands into suitable form, all three data types may be used in Expressions,
Integers are stored as 32 bit twos complement numbers. If an integer overflows it is converted into a real value.
Real values are stored as 64 bit double precision floating point numbers. The exponent can range from -308 to 308, whereas the precision is limited to approximately 15 decimal digits.
Strings are stored as an array of ASCII characters of any length, limited by the Max Memory Request in the Simulation Page of the Simulation Object's Settings. You can view or change the Settings by choosing the Edit / Settings menu item.
System Numerical Attributes now return values that may be either integer, real, or string. You must refer to the definition of the SNA given above to determine the possible data types returned.
Similarly, the system clock is a numeric value which may be either integer or real. This eliminates the need to use large time values to assure fine time granularity.
GPSS function entities are evaluated in double precision form, and the Y values in function follower Statements are, too. This limits precision to approximately 15 decimal digits for the Y value.
SNAs involving GPSS Variable Entities (all types) and those requiring a standard deviation or a division are calculated in double precision floating point format. As a result, these SNAs can experience numeric overflow and/or underflow. Since Expressions are evaluated as double precision floating point numbers, intermediate values are limited to 15 decimal digits of precision and 307 decimal digits of magnitude.
If an overflow occurs during the evaluation of an Expression, an Error Stop occurs. Arithmetic exceptions may occur when divisions, standard deviations, or logarithms are evaluated. An invalid argument to the library functions can also cause an Error Stop.
In the printing of report statistics, when a number in the report is too large for its space in the report, the format of the report is disrupted, but the correct value is printed and the report continues.
Strings are arrays of ASCII characters. Nearly any variable can take on a string value. Arithmetic operators will coerce string values to their numeric equivalent before executing their operations.
String constants are enclosed in double quote marks. When you input a string, such as the file specification in an INCLUDE Command, or a REPORT Command, you must enclose the sequence of ASCII characters in double quotes. Similarly, string values written in Standard Reports and in response to SHOW commands are shown enclosed in double quotes.
To insert a "string within a string" you should use 4 additional double quote characters around the inner string, which is to be sandwiched by 2 pairs of double quotes. When two double quotes together are encountered by the Translator, a single quote is placed in the target string.
Strings are used when you write out simulation results to a Result File. Strings may also be used to format your own ad hoc simulation reports. The Data Stream Blocks OPEN, CLOSE, READ WRITE, SEEK are available for these purposes, as are the PLUS library procedures of the same names. Strings are also used by the DoCommand library procedure during Experiments. Another use for strings is to create a trace for input to an animation post-processor.
You can create and manipulate strings using the routines in the PLUS Procedure Library. If you want to combine strings, use the Catenate library procedure to join 2, or use the more flexible PolyCatenate to join any number of strings. These, and others, are discussed in detail in Chapter 8.
[Table of Contents]