|
When
For each detectable needs to be specified when
the agent or object can detect a certain fact. There are two
options:
whenever:
This means that the detectable is checked
every time a new fact is asserted in the world and for an
agent also every time a new belief is asserted.
at a specified time:
For the detectable needs to be specified
exactly when the detectable needs to be checked by specifying
a percentage varying from 0 to 100% specifying at what
percentage of the workframe completion the detectable needs
to check the fact set and belief set. These kind of
detectables are only checked once.
< Class Variable >
Agents and objects do not always have the ability to
know in advance about what concepts they can detect facts. To allow these agents
and objects to detect facts about unknown concepts it is possible to define the
class of concepts for which facts need to be detected. On the left hand side of
the detectable condition instead of specifying the name of a variable, parameter
or concept instance (agent, object, conceptual object, or area) the class of
concept is specified in between the brackets < and > such as <BaseGroup>. This
will have the agent or object detect all facts of which the left hand side
concept instance is an instance of the specified class or an instance of any of
its sub classes. The matches can be restricted by specifying a more specific
class in the detectable condition. The agent or object will in that case only
detect those instances that are instances of that class or any of its
subclasses, but not any of its superclasses, even if the superclass defines that
attribute and facts for concept instances of that super class exist. A bracketed
class name may also be used on the right hand side of a detectable condition to
indicate that, in the belief matching step, beliefs may have any right hand side
concept that is an instance of that class or any of its subclasses.
Wildcard
A wildcard character, ‘?’, may be used on the right-hand
side of a detectable condition to indicate that the right-hand side value or
object of a matching fact (in step one) or a matching belief (in step 2) is
ignored. Normally, this may be stated more directly by simply omitting the
operator and right-hand side of a value comparison or by omitting the right-hand
side object of a relation comparison. However, the wildcard may occasionally be
useful in a relation comparison together with ‘is false’ or ‘is unknown’ to
match beliefs with truth values other than ‘true’.
Detect-certainty
The detect-certainty is a number ranging from 0
to 100 and represents the percentage of chance that a fact will
be detected based on the detectable. A detect-certainty of 0%
means that the fact will never be detected and basically means
that the detectable is switched off. A detect-certainty of 100%
means that a fact will always be detected based on the
detectable.
Detectable action
There are 5 different detectable actions
possible:
continue:
Has no effect, only used for having agents
or object detect facts and turn them into beliefs.
impasse:
Impasses the workframe on which the agent
or object is working until the impasse is resolved.
abort:
Terminates the workframe on which the agent
or object is working immediately.
complete:
Terminates the workframe on which the agent
or object is working immediately, but still executes all
remaining consequences defined in the workframe. All
remaining activities are skipped.
end_activity:
This action type is only meaning full when
used with composite activities. Causes the composite activity
on which the agent or object is working to be ended.
Defaults
-
when = whenever
-
dc = 100
-
action = continue
Constraints
-
In the comparison the left hand side
attribute type and the right hand side value-type or
right hand side attribute type of a value-expression must
be the same. Object-attribute-index is considered type compatible with
any type and is permitted on both the left and right hand side. For
detectables the right hand side is ignored and therefore no errors or
warnings will ever be generated even if the right hand side resolves to
a value that is type incompatible with the left hand side. However for
the trigger the right hand side is relevant and if the value resulting
from evaluating the right hand side is type incompatible with the left
hand side a warning is generated in the log and the detectable action is
not executed.
-
In the comparison the left hand side and right
hand side types in a relational expression must match the types as
defined for the relation used in the relational expression.
-
ID.name when used as the left hand side concept
must be the name of a concept class, i.e. a group, class, conceptual
class or area definition.
-
The value of the detect-certainty ranges
from 0 to 100 and represents a percentage.
-
The end-activity action type can only be
used when a detectable is defined in a composite
activity.
|
