CPN Tools

State Space Manual

Last updated: January 2006

University of Aarhus
Department of Computer Science
Aabogade 34
DK-8200 Aarhus N, Denmark
Tel: +45 89 42 56 00
Fax: +45 89 42 56 01

2002 University of Aarhus

 2002 University of Aarhus

Department of Computer Science
Aabogade 34
DK-8200 Aarhus N, Denmark
Tel: +45 89 42 56 00
Fax: +45 89 42 56 01
e-mail: [email protected]

Authors: Kurt Jensen, Sren Christensen and Lars M.

Kristensen.

CPN Tools State Space Manual

CPN Tools
State Space Manual
Table of Contents
Index .......................................................................................................................5

Chapter 1 ................................................................................................................7
Introduction to State Spaces ................................................................................7
The History of the State Space Tool .............................................................7
Example: Dining Philosophers ......................................................................7

Chapter 2 ..............................................................................................................11
How to Use the State Space Tools ....................................................................11
Generation of State Space Code ................................................................11
Details and Limitations (can be skipped in a first reading)................12
Generation of the State Space and Scc Graph ...........................................12
Standard Report..........................................................................................13
Statistics Functions...........................................................................14
State Space to Simulator.............................................................................14
Simulator to State Space.............................................................................15

Chapter 3 ..............................................................................................................17
How to Refer to the Items of a State Space .......................................................17
Nodes, Arcs and Strongly Connected Components ....................................17
Place Instances...........................................................................................17
Transition Instances ....................................................................................18
Markings......................................................................................................19
Binding Elements ........................................................................................19
String Representations................................................................................20
Time Values ................................................................................................22

Chapter 4 ..............................................................................................................23
How to Make Standard Queries ..........................................................................23

CPN Tools State Space Manual

State Space Manual

Reachability Properties............................................................................... 23
Boundedness Properties ............................................................................ 24
Home Properties......................................................................................... 26
Liveness Properties .................................................................................... 28
Fairness Properties .................................................................................... 31

Chapter 5.............................................................................................................. 33
How to Make Your Own Queries ........................................................................ 33
Nodes and Arcs .......................................................................................... 33
Strongly Connected Components............................................................... 34
SearchNodes.............................................................................................. 35
PredNodes and EvalNodes.............................................................. 38
Examples of SearchNodes Calls...................................................... 39
SearchArcs ................................................................................................. 40
Examples of SearchArcs Calls......................................................... 41
SearchSccs ................................................................................................ 41
Examples of SearchSccs Calls ........................................................ 42

Chapter 6.............................................................................................................. 43
How to Change Options...................................................................................... 43
String Representation Options.................................................................... 43
Node and Arc Descriptor Options ............................................................... 44
Options for Calculating a State Space........................................................ 45
Stop Options............................................................................................... 46
Branching Options ...................................................................................... 46
Save Report Options .................................................................................. 47
Reference List ..................................................................................................... 49

CPN Tools State Space Manual

Index

Index
A
AllReachable, 24
arc, 17
Arc, 17
arc descriptor, 21
arc descriptor options, 45
Arcs, 33
ArcsInPath, 33
ArcToBE, 20
ArcToScc, 34
ArcToTI, 20

B
BEsDead, 29
BEsFairness, 31
BEsLive, 29
BEsStrictlyLive, 30
BEToTI, 20
Bind, 19
binding element, 19
boundedness property, 13, 14, 24
branching option, 46

C
Calculate Scc Graph tool, 12
Calculate State Space option, 45
Calculate State Space tool, 12
CalculateOccGraph, 12
CalculateSccGraph, 12
chatty version, 23
combination function, 35
CreationTime, 22

D
DeadMarking, 28
DestNode, 33
dining philosopher, 7

E
Enter State Space tool, 11
EntireGraph, 36
EntireGraphCalculated, 14
EqualsUntimed, 22
EqualUntimed, 22
EvalAllArcs, 41
EvalAllNodes, 38
EvalAllSccs, 42
EvalArcs, 41
EvalNodes, 38

EvalSccs, 42
Evaluate ML tool, 12, 23
evaluation function, 35

F
fairness property, 31
FairnessProperty, 31
FullyProcessed, 34

G
generation of Scc graph, 12
generation of state space, 12
generation of state space code, 11

H
home property, 13, 26
HomeMarking, 26
HomeMarkingExists, 27
HomeSpace, 26

I
InArcs, 33
Initial HomeMarking, 27
InitNode, 17
InitScc, 17
InNodes, 33
Inst, 17

L
ListDeadMarkings, 28
ListDeadTIs, 29
ListFairTIs, 32
ListHomeMarkings, 27
ListHomeScc, 27
ListImpartialTIs, 31
ListJustTIs, 32
ListLiveTIs, 30
liveness property, 14, 28
LowerInteger, 25
LowerMultiSet, 25

M
Mark, 19
marking, 19
MinimalHomeSpace, 26

N
node, 17

CPN Tools State Space Manual

State Space Manual

Node, 17
node descriptor, 21
node descriptor option, 44
NodesInPath, 33
NodeToScc, 34
NoLimit, 36
NoOfArcs, 14
NoOfNodes, 14
NoOfSecs, 14

O
OccurrenceTime, 22
option, 43
OutArcs, 33
OutNodes, 33

P
PI, 17
PI.All, 18
place instance, 17
PredAllArcs, 41
PredAllNodes, 38
PredAllSccs, 42
PredArcs, 41
predicate function, 35
PredNodes, 38
PredSccs, 42
Processed, 34

R
reachability property, 23
Reachable, 23

S
Save Report option, 47
Save Report tool, 13
Scc, 17
SccArcs, 34
SccArcsInPath, 34
SccDestNode, 34
SccGraphCalculated, 14
SccInArcs, 34
SccInNodes, 34
SccListDeadMarkings, 28
SccNodesInPath, 34
SccNoOfArcs, 14
SccNoOfNodes, 14
SccNoOfSecs, 14
SccOutArcs, 34
SccOutNodes, 34

CPN Tools State Space Manual

SccReachable, 24
SccSourceNode, 34
SccTerminal, 34
SccToArcs, 34
SccToNodes, 34
SccTrivial, 35
search area, 35
search limit, 35
SearchAllArcs, 41
SearchAllNodes, 38
SearchAllSccs, 42
SearchArcs, 40
SearchNodes, 35
SearchReachableArcs, 41
SearchReachableNodes, 39
SearchReachableSccs, 42
SearchSccs, 41
Simulator to State Space tool, 15
SourceNode, 33
st_Arc, 20
st_BE, 20
st_Mark, 20
st_Node, 20
st_PI, 20
st_TI, 20
standard report, 13
start value, 35
state space code, 11
State Space to Simulator tool, 14
statistics, 13
statistics functions, 14
stop option, 46
string representation, 20
string representation option, 43
StripTime, 22
strongly connected component, 17

T
Terminal, 33
TI, 18
TI.All, 18
timed state space, 22
TIsDead, 28
TIsFairness, 31
TIsLive, 29
transition instance, 18

U
UpperInteger, 24
UpperMultiSet, 25

Chapter 1
Introduction to State Spaces

The History of the State Space Tool

This manual describes a tool to calculate and analyse state spaces (also called occurrence
graphs, reachability graphs or reachability trees).
The SS tool is integrated with CPN Tools. This means that you can easily switch between
the editor, the simulator, and the SS tool. When a state space node has been found, it can
be inspected in the simulator. This means that you can see the marking directly on the
graphical representation of the CPN model. You can see the enabled transition instances,
investigate their bindings and make simulations. Analogously, when a marking has been
found in the simulator, it can be added to the state space or used as the initial marking for
a new state space.
The SS tool has a large number of built-in standard queries. They can be used to
investigate all the standard properties of a CP-net, such as reachability, boundedness,
home properties, liveness and fairness. In addition to the standard queries there are a
number of powerful search facilities allowing you to formulate your own, non-standard
queries. The standard queries require no programming at all. The non-standard queries
usually require that you write 2-5 lines of quite straightforward ML code.

Example: Dining Philosophers

The basic idea behind state spaces is to make a directed graph with a node for each
reachable marking and an arc for each occurring binding element. An introduction to state
spaces can be found in Sect. 5.1 of [CPN 1] and in Sect. 1.1 of [CPN 2].
In this manual we use the dining philosopher system as our main example. Five Chinese
philosophers are sitting around a circular table. In the middle of the table there is a
delicious dish of rice, and between each pair of philosophers there is a single chopstick.
Each philosopher alternates between thinking and eating. To eat, the philosopher needs
two chopsticks, and he is only allowed to use the two which are situated next to him (on
his left and right side). The sharing of chopsticks prevents two neighbours from eating at
the same time.

CPN Tools State Space Manual

State Space Manual

ph1
cs1

cs2

ph5

ph2

Rice
Dish
cs5

cs3
ph4
cs4

ph3

The philosopher system is modelled by the CP-net shown below. The PH colour set
represents the philosophers, while the CS colour set represents the chopsticks. The
function Chopsticks maps each philosopher into the two chopsticks next to him.

val n = 5;
colset PH = index ph with 1..n;
colset CS = index cs with 1..n;
var p: PH;
fun Chopsticks(ph(i)) =
1`cs(i) ++ 1`cs(if i=n then 1 else i+1);
PH.all()
Think
PH
p
Take
Chopsticks

Chopsticks(p)

p
p

CS.all()
Unused
Chopsticks

Eat
PH

CS

p
Chopsticks(p)
Put Down
Chopsticks

CPN Tools State Space Manual

Introduction to State Spaces

A state space for the dining philosophers is shown below. Each node represents a
reachable marking, while each arc represents the occurrence of a single binding element
leading from the marking of the source node to the marking of the destination node. To
improve readability, we have only shown the detailed contents of some of the markings
and some of the binding elements. It should be noted that all arcs are double arcs, i.e.,
they represent two individual arcs.
Unused: 1`cs(3)
Think: 1`ph(2)++1`ph(3)++1`ph(5)
Eat: 1`ph(1)++1`ph(4)
10
2:2

Take: {p=ph(1)}

Take: {p=ph(4)}
2
3:3

7
2:2

8
2:2
3
3:3

6
3:3
1
5:5

Put: {p=ph(2)}

4
3:3

9
2:2
Unused: 1`cs(1)
Think: 1`ph(1)++1`ph(3)++1`ph(5)
Eat: 1`ph(2)++1`ph(4)

5
3:3

Put: {p=ph(3)}

11
2:2
Unused: 1`cs(5)
Think: 1`ph(2)++1`ph(4)++1`ph(5)
Eat: 1`ph(1)++1`ph(3)

The state space for the dining philosophers is generated in the SS tool in CPN Tools.
Currently, there is no support for drawing state spaces in CPN Tools. It should be noted
that this state space is rather atypical. Most state spaces are much larger. The present
version of the SS tool is able to handle graphs with 20,000-200,000 nodes and 50,0002,000,000 arcs provided that you have sufficient RAM in your machine. Future versions
are expected to be able to handle much larger state spaces.

CPN Tools State Space Manual

Chapter 2
How to Use the
State Space Tools
Tools for calculating a state space, saving a state space report, and transferring states
between the simulator and state space tool are found in the state space tool palette:

More information about using the state space tools can be found in the help pages, for
example in the online help:
http://wiki.daimi.au.dk/cpntools-help/use_state_space_tool.wiki

Generation of State Space Code

Before a state space can be calculated, it is necessary to generate the state space code,
i.e., the ML code which is used to calculate and analyse state spaces. To generate the state
space code the following steps must be performed (in the specified order):
a) Make sure that all transitions, places, and pages in the net have names. The names
are required to be unique and to be alphanumeric ML-identifiers: any sequence of
letters, digits, primes () and underscores (_) starting with a letter.
b) Apply the Enter State Space tool to the net, i.e., select Enter SS in the state
space tool palette, and apply it to one of the pages in the net.
The generation of new nodes progresses in a breadth- first fashion. This means that the
nodes are processed in the order in which they were created. To a certain extent, a depth
first generation can be obtained by using narrow Branching Options (described in
Chapter 6).
For a timed state space the processing order is determined by the creation time (i.e., the
model time at which the individual markings start to exist).

CPN Tools State Space Manual

11

State Space Manual

We propose that you now try to generate the state space code for the dining philosopher
system. To do this use the CPN model called DiningPhilosophers. It is located in the
Sample Nets directory of the CPN Tools distribution.
Details and Limitations (can be skipped in a first reading)
When you make a modification of the CPN diagram, it is necessary to regenerate all the
state space code from scratch. This also means that the state space (if any) is lost. When
the modification is made in the simulator it is sufficient to apply the Enter State Space
tool again.
The state space is calculated for those parts of the net which would participate in a simulation. Please note that, in state spaces, it only makes sense to use code segments in a very
limited fashion, e.g., to initialise a CPN model.
Free variables on output arcs are not allowed unless they are variables of a small colour
set.

Generation of the State Space and Scc Graph

When you have generated the state space code (by following the steps described above),
you are ready to calculate the state space.
If the state space is expected to be small (e.g., with a few hundred nodes and arcs), you
can simply apply the Calculate State Space tool, i.e. select the tool from the state space
tool palette and apply it to one of the pages of the net. Otherwise you may need to change
the Stop Options and/or the Branching Options described in Chap. 6.
Many of the query functions in Chap. 4 use the Scc graph (i.e., the strongly connected
components of the state space). To calculate the Scc graph you must apply the Calculate
Scc Graph tool, i.e. select the tool from the state space tool palette and apply it to one of
the pages in the net.
The state space and the Scc graph can also be calculated by using the Evaluate ML tool
to evaluate the following ML functions, which work exactly as the tools in the state space
tool palette. This can, e.g., be useful if you want to handle exceptions raised by the CPN
model or the Stop options (in Chap. 6):
fun CalculateOccGraph

unit -> unit

fun CalculateSccGraph

unit -> unit

Options for the Calculate State Space tool can be found in Chapter 6.

12

CPN Tools State Space Manual

How to Use the State Space Tools

Standard Report
When you have generated the state space for a CP-net, you can use the Save Report tool
to generate a text file which contains a standard report providing information about:
Statistics (size of state space and Scc graph).

Boundedness Properties (integer and multi-set bounds for place instances).

Home Properties (home markings).

Liveness Properties (dead markings, dead/live transition instances).

Fairness Properties (impartial/fair/just transition instances).

Using the options for the Save Report tool (see Chapter 6), one can specify how much
information is to be put in the report. This is done by choosing one or more of the
possibilities mentioned above. Home and fairness properties can be included in the report
only if the Scc graph has been calculated. For the dining philosopher system the full
standard report looks similar to the report below:
Statistics
-----------------------------------------State Space
Nodes: 11
Arcs:
30
Secs:
1
Status: Full
Scc Graph
Nodes: 1
Arcs:
0
Secs:
0

Boundedness Properties
-----------------------------------------Best Integer Bounds
Upper
Lower
System'Eat 1
2
0
System'Think 1
5
3
System'Unused_Chopsticks 1 5
1
Best Upper Multi-set Bounds
System'Eat 1
1`ph(1)++
1`ph(4)++
System'Think 1
1`ph(1)++
1`ph(4)++
System'Unused_Chopsticks 1
1`cs(1)++
1`cs(4)++
Best Lower Multi-set Bounds
System'Eat 1
System'Think 1
System'Unused_Chopsticks 1

1`ph(2)++ 1`ph(3)++
1`ph(5)
1`ph(2)++ 1`ph(3)++
1`ph(5)
1`cs(2)++ 1`cs(3)++
1`cs(5)
empty
empty
empty

Home Properties
-----------------------------------------Home Markings: All

CPN Tools State Space Manual

13

State Space Manual

Liveness Properties
-----------------------------------------Dead Markings: None
Dead Transitions Instances: None
Live Transitions Instances: All
Fairness Properties
-----------------------------------------System'Put_Down_Chopsticks 1
Impartial
System'Take_Chopsticks 1
Impartial

It is possible to customise the way the system displays place instances and transition
instances (e.g. to replace "System'Eat 1" by "Eat"). This is done by means of the
String Representation Options described in Chap. 6.
Statistics Functions
A number of functions can be used to get information about the size of the state space and
the size of the Scc graph.
The state space will always have at least one node (even if the Calculate State Space
tool has not been used). By convention node number 1 represents the initial marking.
If the state space is partial you may extend it by applying the Calculate State Space tool
once more (perhaps after modifying some of the Stop Options or Branching Options (see
Chapter 6)).
Statistics about the calculation of the state space and Scc graph can be accessed via the
following set of ML functions:
fun NoOfNodes

unit -> int

fun NoOfArcs

unit -> int

fun NoOfSecs

unit -> int

fun EntireGraphCalculated

unit -> bool

fun SccNoOfNodes

unit -> int

fun SccNoOfArcs

unit -> int

fun SccNoOfSecs

unit -> int

fun SccGraphCalculated

unit -> bool

State Space to Simulator

The State Space to Simulator tool (SStoSim) allows you to transfer a state from the state
space to the simulator. This is very useful. You can inspect the marking directly on the
graphical representation of the CP-net. You can see the enabled transition instances,
investigate their bindings and make simulations. The SS node (to be moved) is specified
in the tool option for the State Space to Simulator tool.

14

CPN Tools State Space Manual

How to Use the State Space Tools

Simulator to State Space

The Simulator to State Space tool (SimToSS) allows you to transfer a simulator state to
the state space (where it becomes a node). A green status bubble with a message reports
the node number and whether the state already belonged to the state space.

CPN Tools State Space Manual

15

Chapter 3
How to Refer to the Items
of a State Space
This chapter describes how you can refer to the items of a state space, such as nodes,
place instances, binding elements and markings.

Nodes, Arcs and Strongly Connected Components

We denote nodes and arcs by positive integers while we denote strongly connected
components (Sccs) by negative integers:
type Node = int

(* positive *)

type Arc = int

(* positive *)

type Scc = int

(* negative *)

By convention 1 denotes the initial marking of the state space:

val InitNode = 1:Node
while ~1 (minus one) denotes the Scc to which node 1 belongs:
val InitScc = ~1:Scc

Place Instances
To denote place instances the following ML structure is available:
type Inst = int
con PI.<PageName>'<PlaceName> Inst -> PI.PlaceInst

CPN Tools State Space Manual

17

State Space Manual

For the dining philosophers we use:
PI.System'Think 1
to refer to place Think on the first instance of page System. For the ring network from
Sect. 3.1 of [CPN 1] we use:
PI.Site'PackNo 3
to refer to place PackNo on the third instance of the page Site.
You may want to make an alias for place instances frequently referred to, e.g.:
val Eat = PI.System'Eat 1
To denote the set of all place instances, the following notation is available:
PI.All

PI.PlaceInst list

Transition Instances
To denote transition instances the following ML structure is available. It is totally
analogous to PI above:
con TI.<PageName>'<TransName> Inst -> TI.TransInst
For the dining philosophers we use:
TI.System'Take_Chopsticks 1
to refer to transition Take_Chopsticks on the first instance of page System. For the ring
network we use:
TI.Site'Send 3
to refer to transition Send on the third instance of page Site.
To denote the set of all transition instances, the following notation is available:
TI.All

18

CPN Tools State Space Manual

TI.TransInst list

How to Refer to the Items of a State Space

Markings
To inspect the markings of the different place instances the following ML structure is
available:
fun Mark.<PageName>'<PlaceName> Inst -> (Node -> CS ms)
where CS is the colour set of the place instance. For the dining philosophers we use:
Mark.System'Think 1 10
to refer to the multi-set of tokens on place Think on the first instance of page System in
the marking M10 (by convention we use Mi to refer to the marking of node i). For the ring
network we use:
Mark.Site'PackNo 3 217
to refer to the marking of place PackNo on the third instance of the page Site in M217. It
should be noted that the Mark function returns a list representation of the multi-set. To
obtain a string representation of the marking of a place the st_Mark functions should be
used (see String Representations below).
For a timed state space the above functions return a timed multi-set (for places with a
timed colour set).

Binding Elements
To denote binding elements the following ML structure is available:
con Bind.<PageName>'<TransName>
Inst * record -> Bind.Elem
where the second argument is a record specifying the binding of the variables of the
transition. The type of this argument depends upon the transition. For the dining
philosophers we use:
Bind.System'Take_Chopsticks (1,{p=ph(3)})
to refer to the binding element where transition Take_Chopsticks on the first instance of
page System has the variable p bound to ph(3). For the ring network we use:
Bind.Site'NewPack (3,{n=2,r=S(1),s=S(3)})
to refer to the binding element where transition NewPack on the third instance of page
Site has the variables n, r and s bound to 2, S(1) and S(3), respectively.

CPN Tools State Space Manual

19

State Space Manual

The first two of the following functions map an arc into its binding element/transition
instance. The third function maps a binding element into its transition instance.
fun ArcToBE

Arc -> Bind.Elem

fun ArcToTI

Arc -> TI.TransInst

fun BEToTI

Bind.Elem -> TI.TransInst

It should be noted that

Bind.<PageName>'<TransName>
is a constructor. This means that it can be used in pattern matches. Examples can be
found in Examples of SearchArcs Calls in Chap. 5.

String Representations
The following functions are used to obtain string representations of nodes, arcs, place
instances, transition instances and binding elements:
fun st_Node

Node -> string

fun st_Arc

Arc -> string

fun st_PI

PI.PlaceInst -> string

fun st_TI

TI.TransInst -> string

fun st_BE

Bind.Elem -> string

Examples:
st_Node (3)
"3"
st_Arc (18)
"18:6->10"
st_PI (PI.System'Eat 1)
"System'Eat 1"
st_TI (ArcToTI(18)) "System'Take_Chopsticks 1"
st_BE (ArcToBE(18)) "System'Take_Chopsticks 1:{p=ph(4)}"
To produce string representations of the markings of place instances the following ML
structure is provided:
fun st_Mark.<PageName>'<PlaceName>
Inst -> (Node -> string)
Example (the produced string ends with a carriage return):
st_Mark.System'Eat 1 10 "System'Eat 1: 1`ph(1)++1`ph(4)"

20

CPN Tools State Space Manual

How to Refer to the Items of a State Space

The string representations produced by the st-functions can be modified by means of the
String Representation Options in Chap. 6. It is, e.g., possible to get the following more
compact representations (in which place instances and page names are omitted):
st_PI (PI.System'Eat 1)
st_TI (ArcToTI(18))
st_BE (ArcToBE(18))

"Eat"
"Take_Chopsticks"
"Take_Chopsticks: {p=ph(4)}"

Analogously, it is possible to get a compact version of st_Mark (for empty markings the
result is the empty string):
st_Mark.System'Eat 1 10
st_Mark.System'Eat 1 1

"Eat: 1`ph(1)++1`ph(4)"
""

A node descriptor is a string representation of the information associated with a node in

a state space. By default, a node descriptor is a string representation of the node number
and the entire net marking that corresponds to the node in the state space. A node
descriptor can be obtained using the following function:
fun NodeDescriptor

Node -> string

The NodeDescriptor function uses the st_Mark functions for each of the place instances
in the net. Example (the print function prints the string argument):
print(NodeDescriptor 7)
7:
System'Think 1: 1`ph(2)++1`ph(3)++1`ph(5)
System'Eat 1: 1`ph(1)++1`ph(4)
System'Unused_Chopsticks 1: 1`cs(3)
An arc descriptor is a string representation of the information associated with an arc in a
state space. An arc descriptor can be obtained using the following function:
fun ArcDescriptor

Arc -> string

By default, an arc descriptor is a string showing the source and destination nodes of the
arc together with the corresponding binding element. Example
ArcDescriptor 11
"11:3->10 System'Take_Chopsticks 1: {p=ph(2)}"

CPN Tools State Space Manual

21

State Space Manual

Here the arc descriptor for arc 11 indicates that the arc goes from node 3 to node 10, and
the binding element is the Take_Chopsticks transition on the first instance of the page
System where the variable p is bound to ph(2).

Time Values
The following functions can only be used for timed state spaces.
Each node has a time value denoting the model time at which the marking started to
exist:
fun CreationTime

Node -> IntInf.int

Analogously, each arc has a time value denoting the model time at which the binding
element occurred:
fun OccurrenceTime

Arc -> IntInf.int

The following function maps a timed multi-set into an untimed multi-set:

fun StripTime

'a tms -> 'a ms

The following function tells whether the markings of the two specified nodes are identical
when time stamps are ignored:
fun EqualUntimed

Node * Node -> bool

The following function maps a node into all those nodes which have the same marking
when time stamps are ignored:
fun EqualsUntimed

Node -> Node list

Most state spaces with time will be non-cyclic (since all nodes in a cycle must have the
same creation time).

22

CPN Tools State Space Manual

Chapter 4
How to Make Standard Queries
This chapter explains how to perform standard queries to investigate the properties of a
CPN model. It is, e.g., possible to investigate the reachability, boundedness, home,
liveness and fairness properties defined in [CPN 1]. Many of the query functions return
results which already are included in the standard report described in Chap. 2.
The query functions are typically written in auxiliary texts alone or as part of a larger
ML expression. The text is evaluated by means of the Evaluate ML tool.
Some of the functions also have a chatty version which returns the same result as the
ordinary query function. The difference is that the chatty version (sometimes) prints a text
string with a more elaborated explanation of the result. Each chatty query function has the
same name as the corresponding ordinary query function, with a single quote appended to
the end (e.g., Reachable').

Reachability Properties
The query functions for reachability properties are based on Prop 1.12 in [CPN 2].
fun Reachable

Node * Node -> bool

fun SccReachable

Node * Node -> bool

fun AllReachable

unit -> bool

Reachable determines whether there exists an occurrence sequence from the marking
of the first node to the marking of the second. This is done by investigating whether the
state space contains a directed path from the first node to the second. For the dining
philosopher system:
Reachable (5,3)
returns true. This tells us that there exists an occurrence sequence from the marking M5
(of node 5) to the marking M3 (of node 3). The function also has a chatty version:
Reachable' (5,3)
which returns the same result together with the explanation:

CPN Tools State Space Manual

23

State Space Manual

"A path from node 5 to node 3 is: [5,9,3]"
This tells us that there exists an occurrence sequence containing the markings M5, M9
and M3 (in that order). The path is of minimal length.
SccReachable returns the same result as Reachable, but it uses the Scc graph, i.e.,
the strongly connected components. This means that it is faster than Reachable (at
least for state spaces which contain cycles). The function also has a chatty version:
SccReachable' (5,3)
which returns the same result together with the explanation:
"A path from the SCC of node 5 to the SCC of node 3 is:
[~1]"
This tells us that both M5 and M3 belong to the strongly connected component ~1 (i.e. the
strongly connected component of the initial marking).
AllReachable determines whether all the reachable markings are reachable from each
other. This is the case iff there is exactly one strongly connected component. For the
dining philosopher system:
AllReachable ()
returns true.

Boundedness Properties
The query functions for boundedness properties are based on Prop 1.13 in [CPN 2].
fun UpperInteger

(Node -> 'a ms) -> int

fun LowerInteger

(Node -> 'a ms) -> int

fun UpperMultiSet

(Node -> 'a ms) -> 'a ms

fun LowerMultiSet

(Node -> 'a ms) -> 'a ms

UpperInteger uses a specified function F of type:

Node -> 'a ms

24

CPN Tools State Space Manual

How to Make Standard Queries

to calculate an integer |F(n)|. This is done for each node n in the state space, and the
maximum of the calculated integers is returned. For the dining philosopher system:
UpperInteger (Mark.System'Eat 1)
calculates the maximal number of tokens on place Eat on the first instance of page
System. The result is 2, and this tells us that at most two philosophers can eat at the same
time.
LowerInteger is analogous to UpperInteger, but returns the minimal value of the
integers |F(n)|. For the dining philosopher system:
LowerInteger (Mark.System'Think 1)
calculates the minimal number of tokens on place Think on the first instance of page
System. The result is 3, and this tells us that there always are at least three thinking
philosophers.
UpperMultiSet is analogous to UpperInteger, but it calculates F(n) instead of
|F(n)|. The result is the smallest multi-set which is larger than or equal to all the
calculated multi-sets. For the dining philosopher system:
UpperMultiSet (Mark.System'Eat 1)
returns:
[ph 1,ph 2,ph 3,ph 4,ph 5] : PH ms
which is the CPN-ML representation of the multi-set (the elements of the list are not
necessarily sorted):
1`ph(1)++1`ph(2)++1`ph(3)++1`ph(4)++1`ph(5)
This tells us that each of the five philosophers is able to eat. To obtain the above
representation of the result as a string, evaluate the following ML code:
PH.mkstr_ms (UpperMultiSet(Mark.System'Eat 1))
LowerMultiSet is analogous to UpperInteger, but returns the largest multi-set
which is smaller than or equal to all the calculated multi-sets. For the dining philosopher
system:
LowerMultiSet (Mark.System'Eat 1)
returns the empty multi-set. This tells us that each of the five philosophers is able to think
(because there is a marking in which the philosopher is not eating).

CPN Tools State Space Manual

25

State Space Manual

When the four query functions for boundedness are used for a timed place instance of a
timed CP-net, you can use StripTime to get rid of the time stamps, e.g.:
LowerMultiSet (StripTime o (Mark.System'Eat 1))
For more information on StripTime, see Time Values at the end of Chap. 3.

Home Properties
The query functions for home properties are based on Prop 1.14 in [CPN 2].
fun HomeSpace

Node list -> bool

fun MinimalHomeSpace

unit -> int

fun HomeMarking

Node -> bool

fun ListHomeMarkings

unit -> Node list

fun ListHomeScc

unit -> Scc

fun HomeMarkingExists

unit -> bool

fun InitialHomeMarking

unit -> bool

HomeSpace determines whether the set of markings (specified in the list of nodes) is a
home space, i.e., whether, from each reachable marking, it is possible to reach at least one
of the markings. For the dining philosopher system:
HomeSpace [2,6]
returns true. The function also has a chatty version.
MinimalHomeSpace returns the minimal number of markings which is needed to form
a home space. This is identical to the number of terminal strongly connected components.
For the dining philosopher system:
MinimalHomeSpace ()
returns 1. This function cannot be used for a timed CPN model.
HomeMarking determines whether the marking of the specified node is a home
marking, i.e., whether it can be reached from all reachable markings. This is the case iff
there is exactly one terminal strongly connected component and the specified marking
belongs to that component. For the dining philosopher system:
HomeMarking (6)

26

CPN Tools State Space Manual

How to Make Standard Queries

returns true. The function also has a chatty version.
ListHomeMarkings returns a list with all those nodes that are home markings. For the
dining philosopher system:
ListHomeMarkings ()
returns a list which contains all 11 nodes of the state space. This function cannot be used
for a timed CPN model.
ListHomeScc is similar to ListHomeMarkings, but the result is given in a more
compact way. The result is either a single Scc (and then the home markings are exactly
those markings that belong to the Scc) or the result is zero (and then there are no home
markings). For the dining philosophers:
ListHomeScc ()
returns ~1 (i.e. the Scc to which the initial marking belongs). This tells us that all
reachable markings are home markings. This function cannot be used for a timed CPN
model.
HomeMarkingExists determines whether the CP-net has any home markings. This is
the case iff there is exactly one terminal strongly connected component. For the dining
philosopher system:
HomeMarkingExists ()
returns true. This function cannot be used for a timed CPN model.
Initial HomeMarking determines whether the initial marking of the state space is a
home marking, i.e., whether it can be reached from all reachable markings. This is the
case iff there is exactly one strongly connected component. The result of this function is
identical to the result of AllReachable (defined in Reachability Properties). For the
dining philosopher system:
InitialHomeMarking ()
returns true.

CPN Tools State Space Manual

27

State Space Manual

Liveness Properties
The query functions for liveness properties are based on Prop 1.15 in [CPN 2].
fun DeadMarking

Node -> bool

fun ListDeadMarkings

unit -> Node list

fun SccListDeadMarkings

unit -> Node list

fun TIsDead

TI.TransInst list * Node -> bool

fun BEsDead

Bind.Elem list * Node -> bool

fun ListDeadTIs

unit -> TI.TransInst list

fun TIsLive

TI.TransInst list -> bool

fun BEsLive

Bind.Elem list -> bool

fun BEsStrictlyLive

Bind.Elem list -> bool

fun ListLiveTIs

unit -> TI.TransInst list

DeadMarking determines whether the marking of the specified node is dead, i.e., has
no enabled binding elements. For the dining philosopher system:
DeadMarking (8)
returns false. This tells us that M8 has some enabled binding elements.
ListDeadMarkings returns a list with all those nodes that are dead, i.e., have no
enabled binding elements. For the dining philosopher system:
ListDeadMarkings ()
returns the empty list.
SccListDeadMarkings returns the same result as ListDeadMarkings, but it
uses the Scc graph, i.e., the strongly connected components. This means that it is faster
than ListDeadMarkings (at least for state spaces which contain cycles).
TIsDead determines whether the set of transition instances (specified in the list) is dead
in the marking of the specified node, i.e., whether it is impossible to find an occurrence
sequence which starts in the marking and contains one of the transition instances. For the
dining philosopher system:

28

CPN Tools State Space Manual

How to Make Standard Queries

TIsDead ([TI.System'Take_Chopsticks 1],4)
returns false. This tells us that there exists an occurrence sequence which starts in M4 and
contains an occurrence of transition Take_Chopsticks on the first instance of the page
System. The function also has a chatty version:
TIsDead ([TI.System'Take_Chopsticks 1],4)
which returns the same result together with the explanation:
"A transition instance from the given list is contained
in the SCC: ~1 (which is reachable from the SCC of the
given node)"
BEsDead is analogous to TIsDead except that the argument is a list of binding
elements (instead of transition instances). For the dining philosopher system:
BEsDead ([Bind.System'Take_Chopsticks (1,{p=ph(3)})],4)
returns false. This tells us that there exists an occurrence sequence which starts in M4 and
contains an occurrence of transition Take_Chopsticks on the first instance of page System,
with the variable p bound to ph(3). The function also has a chatty version.
ListDeadTIs returns a list with all those transition instances that are dead, i.e., do not
appear in any occurrence sequence starting from the initial marking of the state space. For
the dining philosopher system:
ListDeadTIs ()
returns the empty list.
TIsLive determines whether the set of transition instances (specified in the list) is live,
i.e., whether, from each reachable marking, it is possible to find an occurrence sequence
which contains one of the transition instances. For the dining philosopher system:
TIsLive [TI.System'Take_Chopsticks 1]
returns true. This tells us that it is impossible to reach a marking such that transition
Take_Chopsticks on the first instance of page System never can occur. The function also
has a chatty version.
BEsLive is analogous to TIsLive except that the argument is a list of binding
elements (instead of transition instances). For the dining philosopher system:
BEsLive [Bind.System'Take_Chopsticks (1,{p=ph(3)})]

CPN Tools State Space Manual

29

State Space Manual

returns true. This tells us that philosopher ph(3) always has a chance to Take_Chopsticks.
He cannot do that in all the reachable markings but it is always possible to choose a
sequence of steps so that this may happen. The function also has a chatty version.
BEsStrictlyLive determines whether the set of binding elements (specified in the
list) is strictly live, i.e., whether each individual element in the list is live. For the dining
philosopher system:
BEsStrictlyLive [
Bind.System'Take_Chopsticks
Bind.System'Take_Chopsticks
Bind.System'Take_Chopsticks
Bind.System'Take_Chopsticks
Bind.System'Take_Chopsticks

(1,{p=ph(1)}),
(1,{p=ph(2)}),
(1,{p=ph(3)}),
(1,{p=ph(4)}),
(1,{p=ph(5)})]

returns true. This tells us that each philosopher always has a chance to Take_Chopsticks.
He cannot do that in all the reachable markings but it is always possible to choose a
sequence of steps so that this may happen.
ListLiveTIs returns a list with all those transition instances that are live. For the
dining philosopher system:
ListLiveTIs ()
returns:
[TI.System'Put_Down_Chopsticks 1,
TI.System'Take_Chopsticks 1]
This tells us that it is impossible to reach a marking such that one of the transition
instances never can occur.

30

CPN Tools State Space Manual

How to Make Standard Queries

Fairness Properties
The query functions for fairness properties are based on Prop 1.16 in [CPN 2].
fun TIsFairness

TI.TransInst list ->

FairnessProperty
fun BEsFairness Bind.Elem list ->
FairnessProperty
fun ListImpartialTIs
unit -> TI.TransInst list
fun ListFairTIs

unit -> TI.TransInst list

fun ListJustTIs

unit -> TI.TransInst list

The type FairnessProperty has the following four elements:

{Impartial, Fair, Just, No_Fairness}.
A definition and explanation of impartial, fairness and justice can be found in Sect. 4.5 of
[CPN 1].
TIsFairness determines whether the set of transition instances (specified in the list) is
impartial, fair or just. For the dining philosopher system:
TIsFairness [TI.System'Take_Chopsticks 1]
returns Impartial. This tells us that we cannot have an infinite occurrence sequence
unless transition Take_Chopsticks on the first instance of page System continues to occur.
BEsFairness is analogous to TIsFairness except that the argument is a list of
binding elements (instead of transition instances). For the dining philosopher system:
BEsFairness[Bind.System'Take_Chopsticks (1,{p=ph(3)})]
returns No_Fairness. This tells us that it is possible to have an infinite occurrence
sequence (starting from a reachable marking) in which philosopher three never takes his
chopsticks.
ListImpartialTIs returns a list with all those transition instances that are impartial.
For the dining philosopher system:
ListImpartialTIs ()
returns the list:

CPN Tools State Space Manual

31

State Space Manual

[TI.System'Put_Down_Chopsticks 1,
TI.System'Take_Chopsticks 1]
This tells us that all infinite occurrence sequences (starting from the initial marking)
contains an infinite number of both transition instances.
ListFairTIs and ListJustTIs are analogous to ListImpartialTIs except
that they return all those transition instances that are fair and just, respectively.
Impartiality implies fairness which in turn implies justice. Hence, we always have:
ListImpartialTIs()
ListFairTIs()

ListJustTIs()

32

CPN Tools State Space Manual

Chapter 5
How to Make Your Own Queries
This chapter describes how you can make your own non-standard queries by writing
some simple ML functions. First we introduce a number of functions to inspect the
structure of a state space and an Scc graph. Then we describe three search functions by
which you can traverse the nodes, arcs and strongly connected components of a state
space.

Nodes and Arcs

The following functions allow you to move between adjacent nodes and arcs of the
state space:
fun SourceNode

Arc -> Node

fun DestNode

Arc -> Node

fun Arcs

Node * Node -> Arc list

fun InNodes

Node -> Node list

fun OutNodes

Node -> Node list

fun InArcs

Node -> Arc list

fun OutArcs

Node -> Arc list

The following function tells whether a node is terminal (i.e., have no outgoing arcs). The
result is identical to the result of DeadMarking (in Chap. 4).
fun Terminal

Node -> bool

The following functions return the nodes/arcs in one of the shortest paths between the
two specified nodes. If no path exists the exception NoPathExists is raised:
fun NodesInPath

Node * Node -> Node list

fun ArcsInPath

Node * Node -> Arc list

By definition we always have:

NodesInPath (n,n) = [n]
ArcsInPath (n,n) = []

CPN Tools State Space Manual

33

State Space Manual

The following functions determine to which extent a node has been processed. The
Branching Options (in Chap.6) allow you to specify that a node can be processed without
calculating all the immediate successors. The second function checks whether this is the
case:
fun Processed

Node -> bool

fun FullyProcessed

Node -> bool

Strongly Connected Components

Each node of an Scc graph is a strongly connected component while each arc of an Scc
graph is an ordinary state space arc. We have an Scc arc for each state space arc that
starts in one Scc and ends in another.
The following functions allow you to move between strongly connected components
and their nodes/arcs. The first function maps a state space node into the Scc to which it
belongs. The second function maps a state space arc into the Scc from which it starts (i.e.,
the Scc to which the source node belongs). The third function maps an Scc node into the
state space nodes that belong to the Scc. Finally, the fourth function maps an Scc node
into the state space arcs that start in the Scc (while we don't care where the arcs end):
fun NodeToScc

Node -> Scc

fun ArcToScc

Arc -> Scc

fun SccToNodes

Scc -> Node list

fun SccToArcs

Scc -> Arc list

The following functions (for Scc graphs) are analogous to the functions defined in Nodes
and Arcs (at the beginning of this chapter). Hence they have the same names prefixed
with Scc.

34

fun SccSourceNode

Arc -> Scc

fun SccDestNode

Arc -> Scc

fun SccArcs

Scc * Scc -> Arc list

fun SccInNodes

Scc -> Scc list

fun SccOutNodes

Scc -> Scc list

fun SccInArcs

Scc -> Arc list

fun SccOutArcs

Scc -> Arc list

fun SccTerminal

Scc -> bool

fun SccNodesInPath

Scc * Scc -> Scc list

fun SccArcsInPath

Scc * Scc -> Arc list

CPN Tools State Space Manual

How to Make Your Own Queries

The following function tells whether a strongly connected component is trivial (i.e.,
consists of a single node and no arcs):
fun SccTrivial

Scc -> bool

SearchNodes
The function SearchNodes traverses the nodes of the state space. At each node some
specified calculation is performed and the results of these calculations are combined, in
some specified way, to form the final result.
SearchNodes takes six different arguments and by varying these arguments it is
possible to specify a lot of different queries, e.g., many of the standard queries from
Chap. 4. The following description is taken from Sect. 1.7 of [CPN 2]:
Search Area

Predicate
function
Search Limit

Evaluation
function

Start value
Combination
function

This argument specifies the part of the state space

which should be searched. It is often all nodes, but it
may also be any other subset of nodes, e.g., those
belonging to a strongly connected component.
This argument specifies a function. It maps each
node into a boolean value. Those nodes which
evaluate to false are ignored; the others take part in
the further analysis as described below.
This argument specifies an integer. It tells us how
many times the predicate function may evaluate to
true before we terminate the search. The search limit
may be infinite. This means we always search
through the entire search area.
This argument specifies a function. It maps each
node into a value, of some type A. It is important to
notice that the evaluation function is only used at
those nodes (of the search area) for which the
predicate function evaluates to true.
This argument specifies a constant, of some type B.
This argument specifies a function. It maps from
A B into B, and it describes how each individual
result (obtained by the evaluation function) is
combined with the prior results.

SearchNodes works as described by the following Pascal-style pseudo-code. When the

function terminates it returns the value of Result:

CPN Tools State Space Manual

35

State Space Manual

SearchNodes (Area, Pred, Limit, Eval, Start, Comb)
begin
Result := Start; Found := 0
for all n Area do
if Pred(n) then
begin
Result := Comb(Eval(n), Result)
Found := Found + 1
if Found = Limit then stop for-loop
end
end
end
The arguments have the following types:
area

search area

Node list

pred
limit
eval
start
comb

predicate function
search limit
evaluation function
start value
combine function

Node -> bool

int
Node -> 'a
'b
'a * 'b -> 'b

The ML types 'a and 'b are arbitrary and may be identical. The search area is specified
by a list of nodes (if a node is listed twice it will be searched twice). By convention we
use:
val EntireGraph
to denote the set of all nodes in the state space. The search limit is specified by a positive
integer. By convention we use:
val NoLimit
to specify an infinite limit, i.e., that the search continues until the entire search area has
fully been traversed.
The SearchNodes function is a bit complicated. But it is also extremely general and
powerful. As an example, we can use SearchNodes to implement the query function
ListDeadMarkings from Chap.4, i.e., to find all the dead markings. Then we simply
use the following arguments:

36

CPN Tools State Space Manual

How to Make Your Own Queries

Search area

EntireGraph

Predicate function

fun Pred(n) = (length(OutArcs(n)) = 0)

Search limit

10

Evaluation function

fun Eval(n) = n

Start value

[]

Combination function

fun Comb(new,old) = new::old

The predicate function uses the function OutArcs (from Nodes and Arcs at the
beginning of this chapter) to get a list of all the output arcs. If the length of this list is zero
there are no successors, and thus we have a dead marking. The evaluation function maps
a node into itself, i.e., into the unique node number. The combination function adds each
new dead marking to the list of those which we have previously found. With these
arguments SearchNodes returns a list with at most 10 dead markings. If the list is
empty there is no dead marking. If the length is less than 10, the list contains all the dead
markings. The exact ML call looks as follows:
SearchNodes (
EntireGraph,
fn n => (length(OutArcs(n)) = 0),
10,
fn n => n,
[],
op ::)
As a second example, we may use SearchNodes to implement the query function
UpperInteger from Chap. 4, i.e., to find the best upper integer bound for a given place
instance p PI. This is done by using the following arguments:
Search area

EntireGraph

Predicate function

fun Pred(n) = true

Search limit

NoLimit

Evaluation function

fun Eval(n) = |Mark(p)(n)|

Start value

Combination function

max

The exact ML call looks as follows (for the place Eat on the first instance of the page
System):

CPN Tools State Space Manual

37

State Space Manual

SearchNodes(
EntireGraph,
fn _ => true,
NoLimit,
fn n => size (Mark.System'Eat 1 n),
0,
max)

PredNodes and EvalNodes

For convenience we also define some abbreviated forms of SearchNodes where one or
more of the arguments are predefined. The first function searches the specified area and
returns a list of all those nodes that satisfy the specified predicate. The predeclared
function id maps an arbitrary ML value into itself:
fun PredNodes (area, pred, limit) : Node list
= SearchNodes (area, pred, limit, id, [],op ::)
The second function searches the specified area and returns a list of all the calculated
values:
fun EvalNodes (area, eval) : 'b list
= SearchNodes (area, fn _ => true,
NoLimit, eval, [], op ::)
The next three functions are identical to SearchNodes, PredNodes and
EvalNodes, except that they always search the entire state space:
fun SearchAllNodes(pred, eval, start, comb) : 'b
= SearchNodes (EntireGraph, pred,
NoLimit, eval, start, comb)
fun PredAllNodes (pred) : Node list
= PredNodes (EntireGraph, pred, NoLimit)
fun EvalAllNodes (eval) : 'b list
= EvalNodes (EntireGraph, eval)
The final function is identical to SearchNodes, except that the search area consists of
those nodes that are reachable from the node in the first argument:

38

CPN Tools State Space Manual

How to Make Your Own Queries

fun SearchReachableNodes
(node, pred, limit, eval, start, comb) : 'b

Examples of SearchNodes Calls

Two of the query functions from Chaps. 4 and 6 can be implemented as follows:
fun ListDeadMarkings () : Node list
= PredAllNodes Terminal
fun EntireGraphCalculated () : bool
= (PredAllNodes (fn n => not(FullyProcessed n))= [])
If the state space contains unprocessed nodes, it may be desirable to exclude these from
the node list returned by ListDeadMarkings. We then get the following function:
fun ListDeadMarkingsFP () : Node list
= PredAllNodes (fn n => (Terminal n)
andalso (FullyProcessed n))
All nodes in which a particular philosopher is eating can be found as follows (where cf
returns the coefficient of the specified colour in the specified multi-set):
fun Eating (p:PH) : Node list
= PredAllNodes (fn n => cf(p,Mark.System'Eat 1 n) > 0)
The maximal number of simultaneously enabled transition instances can be found as
follows (where remdupl removes duplicates from a list, while map uses the specified
function on all the elements of the specified list):
fun MaxTIEnabled () : int
= SearchAllNodes(
fn _ => true,
fn n => length(remdupl(map ArcToTI(OutArcs n))),
0,
max)
Checking whether there are reachable markings in which two neighbouring philosophers
simultaneously eat, can be done as follows (where next is a function mapping each
philosopher in its successor, ext_col extends a function 'a -> 'b to a function 'a
ms -> 'b ms, while <<= is the less-than-equal operation on multi-sets):

CPN Tools State Space Manual

39

State Space Manual

fun EatingNeighbours () : Node list
= PredAllNodes(fn n =>
let
val Eating = Mark.System'Eat 1 n
in
not(Eating ++
ext_col next Eating <<= PH.all())
end)
Checking whether there are nodes which violate the linear invariant:
M(Unused_Chopsticks) ++ Chopsticks(M(Eat)) = CS.all()
can be done in the following way (where <><> is the operator which checks whether two
multi-sets are different from each other):
fun InvariantViolations () : Node list
= PredAllNodes(
fn n =>

Mark.System'Unused_Chopsticks 1 n ++
ext_ms Chopsticks (Mark.System'Eat 1 n)
<><> CS.all())

SearchArcs
The function SearchArcs traverses the arcs of the state space. At each arc some
specified calculation is performed and the results of these calculations are combined, in
some specified way, to the form the final result.
We define SearchArcs in a way which is totally analogous to SearchNodes. The
arguments have the following types:
area
pred
limit
eval
start
comb

search area
predicate function
search limit
evaluation function
start value
combine function

Arc list
Arc -> bool
int
Arc -> 'a
'b
'a * 'b -> 'b

We define PredArcs, EvalArcs, SearchAllArcs, PredAllArcs,

EvalAllArcs, and SearchReachableArcs analogously to PredNodes,
40

CPN Tools State Space Manual

How to Make Your Own Queries

EvalNodes, SearchAllNodes, PredAllNodes, EvalAllNodes, and
SearchReachableNodes. The latter searches all the arcs which are reachable from
the node specified in the first argument.
Examples of SearchArcs Calls
The following function returns all the arcs where transition Take_Chopsticks occurs on
the first instance of page System with the variable p bound to a specified philosopher:
fun TakeChopsticks (p:PH) : Arc list
= PredAllArcs(
fn a => case ArcToBE a of
Bind.System'Take_Chopsticks (1,{p=p}) => p=p'
| _ => false)
For the ring network, the following function returns all the arcs where transition Send
occurs on some instance of page Site with variables s and r bound to the same value:
fun SendToMyself () : Arc list
= PredAllArcs(
fn a => case ArcToBE a of
Bind.System'Send (1,{s=v1,r=v2,...}) => v1=v2
| _ => false)

SearchSccs
The function SearchSccs traverses the strongly connected components of the state
space. At each strongly connected component some specified calculation is performed
and the results of these calculations are combined, in some specified way, to the form the
final result.
We define SearchSccs in a way which is totally analogous to SearchNodes and
SearchArcs. The arguments have the following types:
area
pred
limit
eval
start
comb

search area
predicate function
search limit
evaluation function
start value
combine function

Scc list
Scc -> bool
int
Scc -> 'a
'b
'a * 'b -> 'b

We define PredSccs, EvalSccs, SearchAllSccs, PredAllSccs,

EvalAllSccs, and SearchReachableSccs analogously to PredNodes,
EvalNodes, SearchAllNodes, PredAllNodes, EvalAllNodes, and

CPN Tools State Space Manual

41

State Space Manual

SearchReachableNodes. The latter searches all the strongly connected components
which are reachable from the Scc specified in the first argument.
Examples of SearchSccs Calls
Two of the query functions from Chap. 4 can be implemented as follows:
fun HomeMarkingExists () : bool
= (length(PredAllSccs SccTerminal) = 1)
fun HomeMarking (n:Node) : bool
= SccTerminal(NodeToScc(n)) andalso
HomeMarkingExists()

42

CPN Tools State Space Manual

Chapter 6
How to Change Options
The SS tool has a number of options. Some options determine how the string
representation functions work, other options determine the way in which the tools in the
state space tool palette work.

String Representation Options

The String Representation Options allow the user to specify how he wants the stfunctions in Chap. 3 to work. As an example, he may determine whether he wants a
transition instance to be represented as:
System'Take_Chopsticks 1

or

Take

The first representation is convenient for a CP-net with many different pages/page
instances, while the second representation is convenient for a system with only one or a
few pages (and only one instance of each page).
The st-functions are used for the standard reports generated by the Save Report tool and
for the contents of SS node/arc descriptors. Hence, the options also influence these things.
However, it should be noted that the string representation options do not influence the
input format of the different ML functions in Chaps. 3-5. This means, e.g., that the user
always has to specify the page and instance of a transition instance even though he may
have decided to omit this information from the text strings created by st_TI.
For each of the st-functions we provide an ML function which specifies how the
individual substrings are put together (e.g., the order and the separators). The options are
changed by the following set of ML functions (the values indicate the system defaults):
OGSet.StringRepOptions'Node(
fn (node) => node)
OGSet.StringRepOptions'Arc(
fn (arc,source,dest) =>
arc ^ ":" ^ source ^ "->" ^ dest)
OGSet.StringRepOptions'PI(

CPN Tools State Space Manual

43

State Space Manual

fn (page,place,inst) =>
page ^ "'" ^ place ^ " " ^ inst)
OGSet.StringRepOptions'TI(
fn (page,trans,inst) =>
page ^ "'" ^ trans ^ " " ^ inst)
OGSet.StringRepOptions'BE(
fn (TI,bind) => TI ^ ": " ^ bind)
OGSet.StringRepOptions'Mark(
fn (PI,mark) => PI ^ ": " ^ mark ^ "\n")
The more compact string representations mentioned in Chap. 3 are obtained by using the
following options:
OGSet.StringRepOptions'PI(fn (page,place,inst) => place)
OGSet.StringRepOptions'TI(fn (page,trans,inst) => trans)
OGSet.StringRepOptions'BE(
fn (TI,bind) => TI ^ ": " ^ bind)")
OGSet.StringRepOptions'Mark(
fn (PI,mark) =>
if mark="empty" orelse mark="tempty"
then ""
else PI ^ ": " ^ mark ^ "\n")

Node and Arc Descriptor Options

The Node Descriptor Options determine the contents of the SS node descriptors (see
Chapter 3). They are changed by the following ML function (the value indicates the
system defaults for the dining philosopher system):
OGSet.NodeDescriptorOptions(
fn n =>
(st_Node n)^"\n"^
(st_Mark.System'Unused_Chopsticks 1 n)^

44

CPN Tools State Space Manual

How to Change Options

(st_Mark.System'Think 1 n)^
(st_Mark.System'Eat 1 n))
You may replace the default by any other ML function of type:
Node -> string
In this way it is possible to obtain a compact representation of a complex marking. As an
example it is possible to omit the marking of some place instances or only show the
number of tokens (ignoring the token colours).
The Arc Descriptor Options determine the contents of the SS arc descriptors. They are
changed by the following ML function (the value indicates the system defaults for the
dining philosopher system):
OGSet.ArcDescriptorOptions(
fn (a:Arc):string =>
(st_Arc a)^"\n"^
(st_BE(ArcToBE a)))
You may replace the default by any other ML function of type:
Arc -> string
In this way it is possible to obtain a compact representation of a complex binding
element.

Options for Calculating a State Space

There are a number of options for determining how a state space is calculated. There are
stop options and branching options, both of which are described below. Each kind of
option can be changed by modifying the appropriate option for the Calculate State Space
(CalcSS) tool in the index of CPN Tools. The figure below shows the stop options and
the branching options for calculating a state space. The first four options are stop options,
and the last three options are branching options.

CPN Tools State Space Manual

45

State Space Manual

Information about changing options for the Calculate State Space tool can be found in
the help page for the tool, for example in the online help:
http://wiki.daimi.au.dk/cpntools-help/calculate_state_space.wiki

Stop Options
The Stop Options allow you to determine when the calculation of a state space stops. This
happens when all nodes have been processed or when one of the stop options becomes
satisfied. The stop options can be changed by modifying tool options for the Calculate
State Space tool in the index of CPN Tools. The stop options can also be changed by the
following ML function (the values indicate the system defaults):
OGSet.StopOptions{
Nodes = NoLimit,
Arcs = NoLimit,
Secs = 300,
Predicate = fn _ => false}
The first three arguments specify the maximal number of nodes, arcs and seconds. By
convention, zero indicates NoLimit (i.e., that the corresponding stop option is inactive).
All counts are reset to zero whenever you call apply the Calculate State Space tool. This
means that you can extend a state space without changing the Stop Options. The fourth
argument specifies a predicate function which is evaluated after the calculation of the
successors:
Node -> bool
If the predicate evaluates to true the calculation of the state space will be stopped. This
can, e.g., be used to stop when a dead marking is found:
fn n => Terminal n
When a Stop Option has been met, the exception StopOptionSatisfied is raised
(this can, e.g., be seen if the state space is generated by means of the
CalculateOccGraph function described in Chap. 2). Furthermore a message is printed
with details about the activated stop option.
Warning: It is impossible to stop a run-away state space generation in a graceful way.
Hence, it is important that the Stop Options in have some sensible values

Branching Options
The Branching Options allow you to specify that, under certain circumstances, the SS
tool need not calculate all the successors of a node. The node is then said to be only
partially processed. The options can be changed by modifying the appropriate options for

46

CPN Tools State Space Manual

How to Change Options

the Calculate State Space tool. They can also be changed by evaluating the following
ML function (the values indicate the system defaults):
OGSet.BranchingOptions{
TransInsts = NoLimit,
Bindings = NoLimit,
Predicate = fn _ => true}
The first argument specifies the maximal number of enabled transition instances to be
used to find successor markings (for each node). Analogously, the second argument
specifies the maximal number of enabled bindings to be used (for each enabled transition
instance). By convention zero indicates NoLimit. All counts are reset to zero whenever
you apply the Calculate State Space tool. This means that you can extend the number of
calculated successor markings without changing the Branching Options.The third
argument specifies a predicate function which is evaluated before the calculation of the
successors:
Node -> bool
If the predicate evaluates to false no successors are calculated.
Nodes which are processed, without calculating all successors, are marked as partially
processed. When you add to an existing state space, some of the partially processed nodes
may become fully processed.
The generation of new nodes progresses in a breadth- first fashion. This means that the
nodes are being processed in the order in which they were created. To a certain extent, a
depth first generation can be obtained by using narrow Branching Options.
For a timed state space the processing order is determined by the creation time (i.e., the
model time at which the individual markings start to exist).

Save Report Options

As mentioned in Chapter 2, the standard state space report provides information about:

Statistics (size of state space and Scc graph).

Boundedness Properties (integer and multi-set bounds for place instances).

Home Properties (home markings).

Liveness Properties (dead markings, dead/live transition instances).

Fairness Properties (impartial/fair/just transition instances).

CPN Tools State Space Manual

47

State Space Manual

The tool options for the Save Report tool determine which kind of information will be
included in a state space report. The figure below shows the tool options for the Save
Report tool:

48

CPN Tools State Space Manual

Reference List
[CPN 1]

K. Jensen: Coloured Petri Nets. Basic Concepts, Analysis Methods and

Practical Use. Volume 1, Basic Concepts. Monographs in Theoretical
Computer Science, Springer-Verlag, 1992. ISBN: 3-540-60943-1.

[CPN 2]

K. Jensen: Coloured Petri Nets. Basic Concepts, Analysis Methods and

Practical Use. Volume 2, Analysis Methods. Monographs in Theoretical
Computer Science, Springer-Verlag, 1994. ISBN: 3-540-58276-2

CPN Tools State Space Manual

49