E. Baralis et al., COMPILE-TIME AND RUNTIME ANALYSIS OF ACTIVE BEHAVIORS, IEEE transactions on knowledge and data engineering, 10(3), 1998, pp. 353-370
Citations number
39
Categorie Soggetti
Computer Science Artificial Intelligence","Computer Science Information Systems","Engineering, Eletrical & Electronic","Computer Science Artificial Intelligence","Computer Science Information Systems
Active rules may interact in complex and sometimes unpredictable ways,
thus possibly yielding infinite rule executions by triggering each ot
her indefinitely. This paper presents analysis techniques focused on d
etecting termination of rule execution. We describe an approach which
combines static analysis of a rule set at compile-time and detection o
f endless loops during rule processing at runtime. The compile-time an
alysis technique is based on the distinction between mutual triggering
and mutual activation of rules. This distinction motivates the introd
uction of two graphs defining rule interaction, called Triggering and
Activation Graphs, respectively. This analysis technique allows us to
identify reactive behaviors which are guaranteed to terminate and reac
tive behaviors which may lead to infinite rule processing. When termin
ation cannot be guaranteed at compile-time, it is crucial to detect in
finite rule executions at runtime. We propose a technique for identify
ing loops which is based on recognizing that a given situation has alr
eady occurred in the past and, therefore, will occur an infinite numbe
r of times in the future. This technique is potentially Very expensive
, therefore, we explain how it can be implemented in practice with lim
ited computational effort. A particular use of this technique allows u
s to develop cycle monitors, which check that critical rule sequences,
detected at compile time, do not repeat forever. We bridge compile-ti
me analysis to runtime monitoring by showing techniques, based on the
result of rule analysis, for the identification of rule sets that can
be independently monitored and for the optimal selection of cycle moni
tors.