TY - JOUR
T1 - Querying Probabilistic Business Processes for Sub-Flows
AU - Deutch, Daniel
PY - 2013/4
Y1 - 2013/4
N2 - A Business Process (BP for short) consists of a set of activities which, combined in a flow, achieve some business goal. A given BP may have a large, possibly infinite, number of possible execution flows (EX-flows for short), each having some probability to occur at run time. This paper studies query evaluation over such probabilistic BPs. We focus on two important classes of queries, namely boolean queries that compute the probability that a random EX-flow of a BP satisfies a given property, and projection queries focusing on portions of EX-flows that are of interest to the user. For the latter queries the answer consists of the top-k instances of these portions that are most likely to occur at run-time. We study the complexity of query evaluation for both kinds of queries, showing in particular that projection queries may be harder to evaluate than boolean queries. We present a picture of which combinations of BP classes and query features lead to PTIME algorithms and which to NP-hard or infeasible problems.
AB - A Business Process (BP for short) consists of a set of activities which, combined in a flow, achieve some business goal. A given BP may have a large, possibly infinite, number of possible execution flows (EX-flows for short), each having some probability to occur at run time. This paper studies query evaluation over such probabilistic BPs. We focus on two important classes of queries, namely boolean queries that compute the probability that a random EX-flow of a BP satisfies a given property, and projection queries focusing on portions of EX-flows that are of interest to the user. For the latter queries the answer consists of the top-k instances of these portions that are most likely to occur at run-time. We study the complexity of query evaluation for both kinds of queries, showing in particular that projection queries may be harder to evaluate than boolean queries. We present a picture of which combinations of BP classes and query features lead to PTIME algorithms and which to NP-hard or infeasible problems.
KW - Business processes
KW - Probabilistic models
KW - Projection queries
UR - http://www.scopus.com/inward/record.url?scp=84874811176&partnerID=8YFLogxK
U2 - 10.1007/s00224-012-9391-6
DO - 10.1007/s00224-012-9391-6
M3 - ???researchoutput.researchoutputtypes.contributiontojournal.article???
AN - SCOPUS:84874811176
SN - 1432-4350
VL - 52
SP - 367
EP - 402
JO - Theory of Computing Systems
JF - Theory of Computing Systems
IS - 3
ER -