69
Stefan Borgwardt The Complexity Landscape of Temporal Conjunctive Queries Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

  • Upload
    others

  • View
    5

  • Download
    0

Embed Size (px)

Citation preview

Page 1: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Stefan Borgwardt

The Complexity Landscape of TemporalConjunctive QueriesWorkshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Page 2: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Declarative Stream Processing

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 1 of 19

Page 3: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Declarative Stream Processing

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 1 of 19

Page 4: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

High-Level Models

Abstract entities & relations

Discretized input signals

Ontology

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 2 of 19

Page 5: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

High-Level Models

Abstract entities & relations

Discretized input signals

Ontology

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 2 of 19

Page 6: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

High-Level Models

Abstract entities & relations

Discretized input signals

Ontology

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 2 of 19

Page 7: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

In This Talk� Description Logic / OWL Ontologies

� Temporal Logic Queries

� Computational Complexity

AC0 ⊆ NC1 ⊆ P ⊆ NP ⊆ PSPACE ⊆ EXP ⊆ NEXP ⊆ 2EXP

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 3 of 19

Page 8: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

In This Talk� Description Logic / OWL Ontologies

� Temporal Logic Queries

� Computational Complexity

AC0 ⊆ NC1 ⊆ P ⊆ NP ⊆ PSPACE ⊆ EXP ⊆ NEXP ⊆ 2EXP

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 3 of 19

Page 9: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

In This Talk� Description Logic / OWL Ontologies

� Temporal Logic Queries

� Computational Complexity

AC0 ⊆ NC1 ⊆ P ⊆ NP ⊆ PSPACE ⊆ EXP ⊆ NEXP ⊆ 2EXP

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 3 of 19

Page 10: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Description Logic OntologiesConcept names Drone, Component, SensorRole names hasModel, isPartOfIndividual names drone-AB100,myDrone

Concepts Droneu ∃hasSensor.InfraredSensorAssertions drone-AB100 : DeliveryDrone,

(drone-AB100, ErnstReuterPlatz) : hasLocationConcept inclusions DeliveryDrone v Drone,InfraredSensor v ∃measures.Heat

Subclass of

Subclass of

Subclass of

Subclass of

Subclass of

hasCategory

Subclass of

isPartOf(transitive)

hasPart(transitive)

Subclass of

isLandingSystemOf

hasLandingSystemhasSensor

isSensorOf

isDirectPartOf

successorOf

Subclass of

Subclass of

hasModel

isModelOfisFuselageOf

hasFuselage

hasDirectPart

isVariantOf

Subclass of

predecessorOf

AirCraftComp...

Sensor3

LandingSystem

Fuselage

HardwareCo...

Component

ProductModel

SpatialThing(external)

Drone,UAS, UAV

LiftGenerator...

GroundStatio...

Category2

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 4 of 19

Page 11: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Description Logic OntologiesConcept names Drone, Component, SensorRole names hasModel, isPartOfIndividual names drone-AB100,myDroneConcepts Droneu ∃hasSensor.InfraredSensor

Assertions drone-AB100 : DeliveryDrone,(drone-AB100, ErnstReuterPlatz) : hasLocation

Concept inclusions DeliveryDrone v Drone,InfraredSensor v ∃measures.Heat

Subclass of

Subclass of

Subclass of

Subclass of

Subclass of

hasCategory

Subclass of

isPartOf(transitive)

hasPart(transitive)

Subclass of

isLandingSystemOf

hasLandingSystemhasSensor

isSensorOf

isDirectPartOf

successorOf

Subclass of

Subclass of

hasModel

isModelOfisFuselageOf

hasFuselage

hasDirectPart

isVariantOf

Subclass of

predecessorOf

AirCraftComp...

Sensor3

LandingSystem

Fuselage

HardwareCo...

Component

ProductModel

SpatialThing(external)

Drone,UAS, UAV

LiftGenerator...

GroundStatio...

Category2

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 4 of 19

Page 12: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Description Logic OntologiesConcept names Drone, Component, SensorRole names hasModel, isPartOfIndividual names drone-AB100,myDroneConcepts Droneu ∃hasSensor.InfraredSensorAssertions drone-AB100 : DeliveryDrone,

(drone-AB100, ErnstReuterPlatz) : hasLocation

Concept inclusions DeliveryDrone v Drone,InfraredSensor v ∃measures.Heat

Subclass of

Subclass of

Subclass of

Subclass of

Subclass of

hasCategory

Subclass of

isPartOf(transitive)

hasPart(transitive)

Subclass of

isLandingSystemOf

hasLandingSystemhasSensor

isSensorOf

isDirectPartOf

successorOf

Subclass of

Subclass of

hasModel

isModelOfisFuselageOf

hasFuselage

hasDirectPart

isVariantOf

Subclass of

predecessorOf

AirCraftComp...

Sensor3

LandingSystem

Fuselage

HardwareCo...

Component

ProductModel

SpatialThing(external)

Drone,UAS, UAV

LiftGenerator...

GroundStatio...

Category2

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 4 of 19

Page 13: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Description Logic OntologiesConcept names Drone, Component, SensorRole names hasModel, isPartOfIndividual names drone-AB100,myDroneConcepts Droneu ∃hasSensor.InfraredSensorAssertions drone-AB100 : DeliveryDrone,

(drone-AB100, ErnstReuterPlatz) : hasLocationConcept inclusions DeliveryDrone v Drone,InfraredSensor v ∃measures.Heat

Subclass of

Subclass of

Subclass of

Subclass of

Subclass of

hasCategory

Subclass of

isPartOf(transitive)

hasPart(transitive)

Subclass of

isLandingSystemOf

hasLandingSystemhasSensor

isSensorOf

isDirectPartOf

successorOf

Subclass of

Subclass of

hasModel

isModelOfisFuselageOf

hasFuselage

hasDirectPart

isVariantOf

Subclass of

predecessorOf

AirCraftComp...

Sensor3

LandingSystem

Fuselage

HardwareCo...

Component

ProductModel

SpatialThing(external)

Drone,UAS, UAV

LiftGenerator...

GroundStatio...

Category2

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 4 of 19

Page 14: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Description Logic OntologiesConcept names Drone, Component, SensorRole names hasModel, isPartOfIndividual names drone-AB100,myDroneConcepts Droneu ∃hasSensor.InfraredSensorAssertions drone-AB100 : DeliveryDrone,

(drone-AB100, ErnstReuterPlatz) : hasLocationConcept inclusions DeliveryDrone v Drone,InfraredSensor v ∃measures.Heat

Subclass of

Subclass of

Subclass of

Subclass of

Subclass of

hasCategory

Subclass of

isPartOf(transitive)

hasPart(transitive)

Subclass of

isLandingSystemOf

hasLandingSystemhasSensor

isSensorOf

isDirectPartOf

successorOf

Subclass of

Subclass of

hasModel

isModelOfisFuselageOf

hasFuselage

hasDirectPart

isVariantOf

Subclass of

predecessorOf

AirCraftComp...

Sensor3

LandingSystem

Fuselage

HardwareCo...

Component

ProductModel

SpatialThing(external)

Drone,UAS, UAV

LiftGenerator...

GroundStatio...

Category2

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 4 of 19

Page 15: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Temporal Ontologies

A0 A1 . . . AnAssertions

Subclass of

Subclass of

Subclass of

Subclass of

Subclass of

hasCategory

Subclass of

isPartOf(transitive)

hasPart(transitive)

Subclass of

isLandingSystemOf

hasLandingSystemhasSensor

isSensorOf

isDirectPartOf

successorOf

Subclass of

Subclass of

hasModel

isModelOfisFuselageOf

hasFuselage

hasDirectPart

isVariantOf

Subclass of

predecessorOf

AirCraftComp...

Sensor3

LandingSystem

Fuselage

HardwareCo...

Component

ProductModel

SpatialThing(external)

Drone,UAS, UAV

LiftGenerator...

GroundStatio...

Category2

Ontology O

?Query Q

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 5 of 19

Page 16: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Temporal Ontologies

A0 A1 . . . AnAssertions

Subclass of

Subclass of

Subclass of

Subclass of

Subclass of

hasCategory

Subclass of

isPartOf(transitive)

hasPart(transitive)

Subclass of

isLandingSystemOf

hasLandingSystemhasSensor

isSensorOf

isDirectPartOf

successorOf

Subclass of

Subclass of

hasModel

isModelOfisFuselageOf

hasFuselage

hasDirectPart

isVariantOf

Subclass of

predecessorOf

AirCraftComp...

Sensor3

LandingSystem

Fuselage

HardwareCo...

Component

ProductModel

SpatialThing(external)

Drone,UAS, UAV

LiftGenerator...

GroundStatio...

Category2

Ontology O

?Query Q

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 5 of 19

Page 17: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Temporal Ontologies

A0 A1 . . . AnAssertions

Subclass of

Subclass of

Subclass of

Subclass of

Subclass of

hasCategory

Subclass of

isPartOf(transitive)

hasPart(transitive)

Subclass of

isLandingSystemOf

hasLandingSystemhasSensor

isSensorOf

isDirectPartOf

successorOf

Subclass of

Subclass of

hasModel

isModelOfisFuselageOf

hasFuselage

hasDirectPart

isVariantOf

Subclass of

predecessorOf

AirCraftComp...

Sensor3

LandingSystem

Fuselage

HardwareCo...

Component

ProductModel

SpatialThing(external)

Drone,UAS, UAV

LiftGenerator...

GroundStatio...

Category2

Ontology O

?Query Q

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 5 of 19

Page 18: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Temporal Conjunctive Queries (TCQs)Conjunctive queries, a subclass of SQL:

q(d) : ∃s, p. Drone(d)∧ hasSensor(d, s)∧measures(s, p)∧ HighAcceleration(p)

. . . combined via linear temporal operators (plus ∧, ∨, ¬):

Q(d) : Drone(d) ∧ ∃e. hasLocation(d, e)∧ ErnstReuterPlatz(e) ∧

♦− ∃f . hasLocation(d, f )∧ Hauptbahnhof(f )

(Baader, Borgwardt, and Lippmann 2013; Baader, Ghilardi, and Lutz 2012)

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 6 of 19

Page 19: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Temporal Conjunctive Queries (TCQs)Conjunctive queries, a subclass of SQL:

q(d) : ∃s, p. Drone(d)∧ hasSensor(d, s)∧measures(s, p)∧ HighAcceleration(p)

. . . combined via linear temporal operators (plus ∧, ∨, ¬):

Q(d) : Drone(d) ∧ ∃e. hasLocation(d, e)∧ ErnstReuterPlatz(e) ∧

♦− ∃f . hasLocation(d, f )∧ Hauptbahnhof(f )

(Baader, Borgwardt, and Lippmann 2013; Baader, Ghilardi, and Lutz 2012)

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 6 of 19

Page 20: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Temporal Concepts� Rigid names have a fixed interpretation: Drone, hasSensor

� Flexible names are allowed to change: hasLocation� No other form of temporal concepts, e.g.

Drone v ♦∃hasLocation.MaintenanceYard

More expressive and expensive

Undecidable in combination with rigid roles

No tools to develop temporal DL ontologies exist

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 7 of 19

Page 21: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Temporal Concepts� Rigid names have a fixed interpretation: Drone, hasSensor� Flexible names are allowed to change: hasLocation

� No other form of temporal concepts, e.g.

Drone v ♦∃hasLocation.MaintenanceYard

More expressive and expensive

Undecidable in combination with rigid roles

No tools to develop temporal DL ontologies exist

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 7 of 19

Page 22: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Temporal Concepts� Rigid names have a fixed interpretation: Drone, hasSensor� Flexible names are allowed to change: hasLocation� No other form of temporal concepts, e.g.

Drone v ♦∃hasLocation.MaintenanceYard

More expressive and expensive

Undecidable in combination with rigid roles

No tools to develop temporal DL ontologies exist

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 7 of 19

Page 23: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Temporal Concepts� Rigid names have a fixed interpretation: Drone, hasSensor� Flexible names are allowed to change: hasLocation� No other form of temporal concepts, e.g.

Drone v ♦∃hasLocation.MaintenanceYard

More expressive and expensive

Undecidable in combination with rigid roles

No tools to develop temporal DL ontologies exist

(Gabbay, Kurucz, Wolter, and Zakharyaschev 2003)

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 7 of 19

Page 24: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Answering TCQsSubclass of

Subclass of

Subclass of

Subclass of

Subclass of

hasCategory

Subclass of

isPartOf(transitive)

hasPart(transitive)

Subclass of

isLandingSystemOf

hasLandingSystemhasSensor

isSensorOf

isDirectPartOf

successorOf

Subclass of

Subclass of

hasModel

isModelOfisFuselageOf

hasFuselage

hasDirectPart

isVariantOf

Subclass of

predecessorOf

AirCraftComp...

Sensor3

LandingSystem

Fuselage

HardwareCo...

Component

ProductModel

SpatialThing(external)

Drone,UAS, UAV

LiftGenerator...

GroundStatio...

Category2

?

Does the query Q follow from the ontology O and observations A0, . . . ,An,taking the rigid names into account?

Do Q, O, and A0, . . . ,An have a common model,respecting the rigid names?

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 8 of 19

Page 25: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Answering TCQsSubclass of

Subclass of

Subclass of

Subclass of

Subclass of

hasCategory

Subclass of

isPartOf(transitive)

hasPart(transitive)

Subclass of

isLandingSystemOf

hasLandingSystemhasSensor

isSensorOf

isDirectPartOf

successorOf

Subclass of

Subclass of

hasModel

isModelOfisFuselageOf

hasFuselage

hasDirectPart

isVariantOf

Subclass of

predecessorOf

AirCraftComp...

Sensor3

LandingSystem

Fuselage

HardwareCo...

Component

ProductModel

SpatialThing(external)

Drone,UAS, UAV

LiftGenerator...

GroundStatio...

Category2

?

Does the query Q follow from the ontology O and observations A0, . . . ,An,taking the rigid names into account?

Do Q, O, and A0, . . . ,An have a common model,respecting the rigid names?

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 8 of 19

Page 26: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Complexity LandscapeDescription

logic

Data complexity Combined complexity

0 1 2 3 0 1 2 3

DL-Lite AC0

NC1

NP PSPACE

EL P CONP NP PSPACE CONEXP

ALC CONP EXP CONEXP 2EXP

SHIQ CONP ≤EXP 2EXP

SHO[Q|I] CONP ≤ ? ≤ EXP 2EXP

ALCHOIQ CONP ≤ ? ≤ dec. CO2NEXP ≤ ? ≤ dec.SROIQ CONP ≤ ? CO2NEXP ≤ ?

0 atemporal (CQs) 1 no rigid names 2 rigid concept names 3 rigid role names

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 9 of 19

Page 27: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Complexity LandscapeDescription

logic

Data complexity Combined complexity

0 1 2 3 0 1 2 3

DL-Lite AC0

NC1

NP PSPACE

EL P CONP NP PSPACE CONEXP

ALC CONP EXP CONEXP 2EXP

SHIQ CONP ≤EXP 2EXP

SHO[Q|I] CONP ≤ ? ≤ EXP 2EXP

ALCHOIQ CONP ≤ ? ≤ dec. CO2NEXP ≤ ? ≤ dec.SROIQ CONP ≤ ? CO2NEXP ≤ ?

0 atemporal (CQs) 1 no rigid names 2 rigid concept names 3 rigid role names

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 9 of 19

Page 28: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Complexity LandscapeDescription

logic

Data complexity Combined complexity

0 1 2 3 0 1 2 3

DL-Lite AC0

NC1

NP PSPACE

EL P CONP NP PSPACE CONEXP

ALC CONP EXP CONEXP 2EXP

SHIQ CONP ≤EXP 2EXP

SHO[Q|I] CONP ≤ ? ≤ EXP 2EXP

ALCHOIQ CONP ≤ ? ≤ dec. CO2NEXP ≤ ? ≤ dec.SROIQ CONP ≤ ? CO2NEXP ≤ ?

0 atemporal (CQs) 1 no rigid names 2 rigid concept names 3 rigid role names

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 9 of 19

Page 29: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Complexity LandscapeDescription

logic

Data complexity Combined complexity

0 1 2 3 0 1 2 3

DL-Lite AC0

NC1

NP PSPACE

EL P CONP NP PSPACE CONEXP

ALC CONP EXP CONEXP 2EXP

SHIQ CONP ≤EXP 2EXP

SHO[Q|I] CONP ≤ ? ≤ EXP 2EXP

ALCHOIQ CONP ≤ ? ≤ dec. CO2NEXP ≤ ? ≤ dec.SROIQ CONP ≤ ? CO2NEXP ≤ ?

0 atemporal (CQs) 1 no rigid names 2 rigid concept names 3 rigid role names

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 9 of 19

Page 30: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Complexity LandscapeDescription

logic

Data complexity Combined complexity

0 1 2 3 0 1 2 3

DL-Lite AC0

NC1

NP PSPACE

EL P CONP NP PSPACE CONEXP

ALC CONP EXP CONEXP 2EXP

SHIQ CONP ≤EXP 2EXP

SHO[Q|I] CONP ≤ ? ≤ EXP 2EXP

ALCHOIQ CONP ≤ ? ≤ dec. CO2NEXP ≤ ? ≤ dec.SROIQ CONP ≤ ? CO2NEXP ≤ ?

0 atemporal (CQs) 1 no rigid names 2 rigid concept names 3 rigid role names

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 9 of 19

Page 31: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Complexity LandscapeDescription

logic

Data complexity Combined complexity

0 1 2 3 0 1 2 3

DL-Lite AC0

NC1

NP PSPACE

EL P CONP NP PSPACE CONEXP

ALC CONP EXP CONEXP 2EXP

SHIQ CONP ≤EXP 2EXP

SHO[Q|I] CONP ≤ ? ≤ EXP 2EXP

ALCHOIQ CONP ≤ ? ≤ dec. CO2NEXP ≤ ? ≤ dec.

SROIQ CONP ≤ ? CO2NEXP ≤ ?

0 atemporal (CQs) 1 no rigid names 2 rigid concept names 3 rigid role names

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 9 of 19

Page 32: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Complexity LandscapeDescription

logic

Data complexity Combined complexity

0 1 2 3 0 1 2 3

DL-Lite AC0

NC1

NP PSPACE

EL P CONP NP PSPACE CONEXP

ALC CONP EXP CONEXP 2EXP

SHIQ CONP ≤EXP 2EXP

SHO[Q|I] CONP ≤ ? ≤ EXP 2EXP

ALCHOIQ CONP ≤ ? ≤ dec. CO2NEXP ≤ ? ≤ dec.SROIQ CONP ≤ ? CO2NEXP ≤ ?

0 atemporal (CQs) 1 no rigid names 2 rigid concept names 3 rigid role names

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 9 of 19

Page 33: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Propositional Abstraction� To decide satisfiability of Q over O and A0, . . . ,An,replace the inner CQs q1, . . . , qm by propositional variables p1, . . . , pm.

Qp : p1 ∧ p2 ∧ ♦−p3

� possible worldsW = {W1, . . . ,Wk} ⊆ 2{p1,...,pm}, e.g. {p1, p3}� mapping A : {0, . . . , n}→ {1, . . . , k}

WA(0) WA(1) . . . WA(n)

?Qp

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 10 of 19

Page 34: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Propositional Abstraction� To decide satisfiability of Q over O and A0, . . . ,An,replace the inner CQs q1, . . . , qm by propositional variables p1, . . . , pm.

Qp : p1 ∧ p2 ∧ ♦−p3

� possible worldsW = {W1, . . . ,Wk} ⊆ 2{p1,...,pm}, e.g. {p1, p3}

� mapping A : {0, . . . , n}→ {1, . . . , k}

WA(0) WA(1) . . . WA(n)

?Qp

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 10 of 19

Page 35: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Propositional Abstraction� To decide satisfiability of Q over O and A0, . . . ,An,replace the inner CQs q1, . . . , qm by propositional variables p1, . . . , pm.

Qp : p1 ∧ p2 ∧ ♦−p3

� possible worldsW = {W1, . . . ,Wk} ⊆ 2{p1,...,pm}, e.g. {p1, p3}� mapping A : {0, . . . , n}→ {1, . . . , k}

WA(0) WA(1) . . . WA(n)

?Qp

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 10 of 19

Page 36: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Main Approach

WA(0) WA(1) . . . WA(n) tn+1 . . . Qk

?Qp

T-satisfiability

LTL model t0, t1, . . . of Qp with� ti = WA(i) for 0 ≤ i ≤ n� ti ∈W for i > n

P / PSPACE

Subclass of

Subclass of

Subclass of

Subclass of

Subclass of

hasCategory

Subclass of

isPartOf(transitive)

hasPart(transitive)

Subclass of

isLandingSystemOf

hasLandingSystemhasSensor

isSensorOf

isDirectPartOf

successorOf

Subclass of

Subclass of

hasModel

isModelOfisFuselageOf

hasFuselage

hasDirectPart

isVariantOf

Subclass of

predecessorOf

AirCraftComp...

Sensor3

LandingSystem

Fuselage

HardwareCo...

Component

ProductModel

SpatialThing(external)

Drone,UAS, UAV

LiftGenerator...

GroundStatio...

Category2

R-satisfiability

Models of

� QA(i),Ai and O for 0 ≤ i ≤ n� Qi and O for 1 ≤ i ≤ kthat respect the rigid names.

Qi :∧pj∈Wi

qj ∧∧pj /∈Wi

¬qj

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 11 of 19

Page 37: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Main Approach

QA(0),A0 QA(1),A1 . . . QA(n),An Q1 . . . Qk

?Qp

T-satisfiability

LTL model t0, t1, . . . of Qp with� ti = WA(i) for 0 ≤ i ≤ n� ti ∈W for i > n

P / PSPACE

Subclass of

Subclass of

Subclass of

Subclass of

Subclass of

hasCategory

Subclass of

isPartOf(transitive)

hasPart(transitive)

Subclass of

isLandingSystemOf

hasLandingSystemhasSensor

isSensorOf

isDirectPartOf

successorOf

Subclass of

Subclass of

hasModel

isModelOfisFuselageOf

hasFuselage

hasDirectPart

isVariantOf

Subclass of

predecessorOf

AirCraftComp...

Sensor3

LandingSystem

Fuselage

HardwareCo...

Component

ProductModel

SpatialThing(external)

Drone,UAS, UAV

LiftGenerator...

GroundStatio...

Category2

R-satisfiability

Models of

� QA(i),Ai and O for 0 ≤ i ≤ n� Qi and O for 1 ≤ i ≤ kthat respect the rigid names.

Qi :∧pj∈Wi

qj ∧∧pj /∈Wi

¬qj

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 11 of 19

Page 38: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

R-satisfiability in Case 1 (No Rigid Names)

QA(0),A0 QA(1),A1 . . . QA(n),An Q1 . . . Qk

Subclass of

Subclass of

Subclass of

Subclass of

Subclass of

hasCategory

Subclass of

isPartOf(transitive)

hasPart(transitive)

Subclass of

isLandingSystemOf

hasLandingSystemhasSensor

isSensorOf

isDirectPartOf

successorOf

Subclass of

Subclass of

hasModel

isModelOfisFuselageOf

hasFuselage

hasDirectPart

isVariantOf

Subclass of

predecessorOf

AirCraftComp...

Sensor3

LandingSystem

Fuselage

HardwareCo...

Component

ProductModel

SpatialThing(external)

Drone,UAS, UAV

LiftGenerator...

GroundStatio...

Category2

independent atemporal satisfiability checks

“maximal” setW, mapping A

EL P CONP NP PSPACE CONEXP

ALC CONP EXP CONEXP 2EXP

(Baader, Borgwardt, and Lippmann 2013; Borgwardt and Thost 2015b)

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 12 of 19

Page 39: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

R-satisfiability in Case 1 (No Rigid Names)

QA(0),A0 QA(1),A1 . . . QA(n),An Q1 . . . Qk

Subclass of

Subclass of

Subclass of

Subclass of

Subclass of

hasCategory

Subclass of

isPartOf(transitive)

hasPart(transitive)

Subclass of

isLandingSystemOf

hasLandingSystemhasSensor

isSensorOf

isDirectPartOf

successorOf

Subclass of

Subclass of

hasModel

isModelOfisFuselageOf

hasFuselage

hasDirectPart

isVariantOf

Subclass of

predecessorOf

AirCraftComp...

Sensor3

LandingSystem

Fuselage

HardwareCo...

Component

ProductModel

SpatialThing(external)

Drone,UAS, UAV

LiftGenerator...

GroundStatio...

Category2

independent atemporal satisfiability checks

“maximal” setW, mapping A

EL P CONP NP PSPACE CONEXP

ALC CONP EXP CONEXP 2EXP

(Baader, Borgwardt, and Lippmann 2013; Borgwardt and Thost 2015b)

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 12 of 19

Page 40: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

R-satisfiability in Case 1 (No Rigid Names)

QA(0),A0 QA(1),A1 . . . QA(n),An Q1 . . . Qk

Subclass of

Subclass of

Subclass of

Subclass of

Subclass of

hasCategory

Subclass of

isPartOf(transitive)

hasPart(transitive)

Subclass of

isLandingSystemOf

hasLandingSystemhasSensor

isSensorOf

isDirectPartOf

successorOf

Subclass of

Subclass of

hasModel

isModelOfisFuselageOf

hasFuselage

hasDirectPart

isVariantOf

Subclass of

predecessorOf

AirCraftComp...

Sensor3

LandingSystem

Fuselage

HardwareCo...

Component

ProductModel

SpatialThing(external)

Drone,UAS, UAV

LiftGenerator...

GroundStatio...

Category2

independent atemporal satisfiability checks

“maximal” setW, mapping A

EL P CONP NP PSPACE CONEXP

ALC CONP EXP CONEXP 2EXP

(Baader, Borgwardt, and Lippmann 2013; Borgwardt and Thost 2015b)

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 12 of 19

Page 41: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

R-satisfiability in Case 3 (Rigid Role Names)

QA(0),A0 QA(1),A1 . . . QA(n),An Q1 . . . Qk

Subclass of

Subclass of

Subclass of

Subclass of

Subclass of

hasCategory

Subclass of

isPartOf(transitive)

hasPart(transitive)

Subclass of

isLandingSystemOf

hasLandingSystemhasSensor

isSensorOf

isDirectPartOf

successorOf

Subclass of

Subclass of

hasModel

isModelOfisFuselageOf

hasFuselage

hasDirectPart

isVariantOf

Subclass of

predecessorOf

AirCraftComp...

Sensor3

LandingSystem

Fuselage

HardwareCo...

Component

ProductModel

SpatialThing(external)

Drone,UAS, UAV

LiftGenerator...

GroundStatio...

Category2

integrate satisfiability check by renaming the flexible names:

Q(0)A(0),A

(0)0, . . . ,Q(n)

A(n),A(n)n , Q(n+1)

1, . . . ,Q(n+k)

k , O(0), . . . ,O(n+k)

EL P CONP NP PSPACE CONEXP

ALC CONP EXP CONEXP 2EXP

(Baader, Borgwardt, and Lippmann 2013; Borgwardt and Thost 2015b)

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 13 of 19

Page 42: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

R-satisfiability in Case 3 (Rigid Role Names)

QA(0),A0 QA(1),A1 . . . QA(n),An Q1 . . . Qk

Subclass of

Subclass of

Subclass of

Subclass of

Subclass of

hasCategory

Subclass of

isPartOf(transitive)

hasPart(transitive)

Subclass of

isLandingSystemOf

hasLandingSystemhasSensor

isSensorOf

isDirectPartOf

successorOf

Subclass of

Subclass of

hasModel

isModelOfisFuselageOf

hasFuselage

hasDirectPart

isVariantOf

Subclass of

predecessorOf

AirCraftComp...

Sensor3

LandingSystem

Fuselage

HardwareCo...

Component

ProductModel

SpatialThing(external)

Drone,UAS, UAV

LiftGenerator...

GroundStatio...

Category2

integrate satisfiability check by renaming the flexible names:

Q(0)A(0),A

(0)0, . . . ,Q(n)

A(n),A(n)n , Q(n+1)

1, . . . ,Q(n+k)

k , O(0), . . . ,O(n+k)

EL P CONP NP PSPACE CONEXP

ALC CONP EXP CONEXP 2EXP

(Baader, Borgwardt, and Lippmann 2013; Borgwardt and Thost 2015b)

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 13 of 19

Page 43: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

R-satisfiability in Case 3 (Rigid Role Names)

QA(0),A0 QA(1),A1 . . . QA(n),An Q1 . . . Qk

Subclass of

Subclass of

Subclass of

Subclass of

Subclass of

hasCategory

Subclass of

isPartOf(transitive)

hasPart(transitive)

Subclass of

isLandingSystemOf

hasLandingSystemhasSensor

isSensorOf

isDirectPartOf

successorOf

Subclass of

Subclass of

hasModel

isModelOfisFuselageOf

hasFuselage

hasDirectPart

isVariantOf

Subclass of

predecessorOf

AirCraftComp...

Sensor3

LandingSystem

Fuselage

HardwareCo...

Component

ProductModel

SpatialThing(external)

Drone,UAS, UAV

LiftGenerator...

GroundStatio...

Category2

integrate satisfiability check by renaming the flexible names:

Q(0)A(0),A

(0)0, . . . ,Q(n)

A(n),A(n)n , Q(n+1)

1, . . . ,Q(n+k)

k , O(0), . . . ,O(n+k)

EL P CONP NP PSPACE CONEXP

ALC CONP EXP CONEXP 2EXP

(Baader, Borgwardt, and Lippmann 2013; Borgwardt and Thost 2015b)

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 13 of 19

Page 44: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Complexity LandscapeDescription

logic

Data complexity Combined complexity

0 1 2 3 0 1 2 3

DL-Lite AC0

NC1

NP PSPACE

EL P CONP NP PSPACE CONEXP

ALC CONP EXP CONEXP 2EXP

SHIQ CONP ≤EXP 2EXP

SHO[Q|I] CONP ≤ ? ≤ EXP 2EXP

ALCHOIQ CONP ≤ ? ≤ dec. CO2NEXP ≤ ? ≤ dec.SROIQ CONP ≤ ? CO2NEXP ≤ ?

0 atemporal (CQs) 1 no rigid names 2 rigid concept names 3 rigid role names

(Baader, Borgwardt, and Lippmann 2015a,b; Borgwardt and Thost 2015b; Thost 2017)

(Borgwardt and Thost 2015a)

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 14 of 19

Page 45: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Complexity LandscapeDescription

logic

Data complexity Combined complexity

0 1 2 3 0 1 2 3

DL-Lite AC0

NC1

NP PSPACE

EL P CONP NP PSPACE CONEXP

ALC CONP EXP CONEXP 2EXP

SHIQ CONP ≤EXP 2EXP

SHO[Q|I] CONP ≤ ? ≤ EXP 2EXP

ALCHOIQ CONP ≤ ? ≤ dec. CO2NEXP ≤ ? ≤ dec.SROIQ CONP ≤ ? CO2NEXP ≤ ?

0 atemporal (CQs) 1 no rigid names 2 rigid concept names 3 rigid role names

(Baader, Borgwardt, and Lippmann 2015a,b; Borgwardt and Thost 2015b; Thost 2017)

(Borgwardt and Thost 2015a)

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 14 of 19

Page 46: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

DL-Lite AC0

NC1

NP PSPACE

� Guess additional information P� Do not constructW and A explicitly� Combine T-satisfiability and R-satisfiability checks

Q0,A0 Q1,A1 . . . Qn,An Qn+1 . . .

Subclass of

Subclass of

Subclass of

Subclass of

Subclass of

hasCategory

Subclass of

isPartOf(transitive)

hasPart(transitive)

Subclass of

isLandingSystemOf

hasLandingSystemhasSensor

isSensorOf

isDirectPartOf

successorOf

Subclass of

Subclass of

hasModel

isModelOfisFuselageOf

hasFuselage

hasDirectPart

isVariantOf

Subclass of

predecessorOf

AirCraftComp...

Sensor3

LandingSystem

Fuselage

HardwareCo...

Component

ProductModel

SpatialThing(external)

Drone,UAS, UAV

LiftGenerator...

GroundStatio...

Category2

P P P P P P P

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 15 of 19

Page 47: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

DL-Lite AC0

NC1

NP PSPACE

� Guess additional information P� Do not constructW and A explicitly� Combine T-satisfiability and R-satisfiability checks

Q0,A0 Q1,A1 . . . Qn,An Qn+1 . . .

Subclass of

Subclass of

Subclass of

Subclass of

Subclass of

hasCategory

Subclass of

isPartOf(transitive)

hasPart(transitive)

Subclass of

isLandingSystemOf

hasLandingSystemhasSensor

isSensorOf

isDirectPartOf

successorOf

Subclass of

Subclass of

hasModel

isModelOfisFuselageOf

hasFuselage

hasDirectPart

isVariantOf

Subclass of

predecessorOf

AirCraftComp...

Sensor3

LandingSystem

Fuselage

HardwareCo...

Component

ProductModel

SpatialThing(external)

Drone,UAS, UAV

LiftGenerator...

GroundStatio...

Category2

P P P P P P P

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 15 of 19

Page 48: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

DL-Lite AC0

NC1

NP PSPACE

� Guess additional information P� Do not constructW and A explicitly� Combine T-satisfiability and R-satisfiability checks

Q0,A0 Q1,A1 . . . Qn,An Qn+1 . . .

Subclass of

Subclass of

Subclass of

Subclass of

Subclass of

hasCategory

Subclass of

isPartOf(transitive)

hasPart(transitive)

Subclass of

isLandingSystemOf

hasLandingSystemhasSensor

isSensorOf

isDirectPartOf

successorOf

Subclass of

Subclass of

hasModel

isModelOfisFuselageOf

hasFuselage

hasDirectPart

isVariantOf

Subclass of

predecessorOf

AirCraftComp...

Sensor3

LandingSystem

Fuselage

HardwareCo...

Component

ProductModel

SpatialThing(external)

Drone,UAS, UAV

LiftGenerator...

GroundStatio...

Category2

P P P P P P P

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 15 of 19

Page 49: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

DL-Lite AC0

NC1

NP PSPACE

� Guess additional information P� Do not constructW and A explicitly� Combine T-satisfiability and R-satisfiability checks

Q0,A0 Q1,A1 . . . Qn,An Qn+1 . . .

Subclass of

Subclass of

Subclass of

Subclass of

Subclass of

hasCategory

Subclass of

isPartOf(transitive)

hasPart(transitive)

Subclass of

isLandingSystemOf

hasLandingSystemhasSensor

isSensorOf

isDirectPartOf

successorOf

Subclass of

Subclass of

hasModel

isModelOfisFuselageOf

hasFuselage

hasDirectPart

isVariantOf

Subclass of

predecessorOf

AirCraftComp...

Sensor3

LandingSystem

Fuselage

HardwareCo...

Component

ProductModel

SpatialThing(external)

Drone,UAS, UAV

LiftGenerator...

GroundStatio...

Category2

P P P P P P P

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 15 of 19

Page 50: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

DL-Lite AC0

NC1

NP PSPACE

� Guess additional information P� Do not constructW and A explicitly� Combine T-satisfiability and R-satisfiability checks

Q0,A0 Q1,A1 . . . Qn,An Qn+1 . . .

Subclass of

Subclass of

Subclass of

Subclass of

Subclass of

hasCategory

Subclass of

isPartOf(transitive)

hasPart(transitive)

Subclass of

isLandingSystemOf

hasLandingSystemhasSensor

isSensorOf

isDirectPartOf

successorOf

Subclass of

Subclass of

hasModel

isModelOfisFuselageOf

hasFuselage

hasDirectPart

isVariantOf

Subclass of

predecessorOf

AirCraftComp...

Sensor3

LandingSystem

Fuselage

HardwareCo...

Component

ProductModel

SpatialThing(external)

Drone,UAS, UAV

LiftGenerator...

GroundStatio...

Category2

P P P P P P P

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 15 of 19

Page 51: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

DL-Lite AC0

NC1

NP PSPACE

� Guess additional information P� Do not constructW and A explicitly� Combine T-satisfiability and R-satisfiability checks

Q0,A0 Q1,A1 . . . Qn,An Qn+1 . . .

Subclass of

Subclass of

Subclass of

Subclass of

Subclass of

hasCategory

Subclass of

isPartOf(transitive)

hasPart(transitive)

Subclass of

isLandingSystemOf

hasLandingSystemhasSensor

isSensorOf

isDirectPartOf

successorOf

Subclass of

Subclass of

hasModel

isModelOfisFuselageOf

hasFuselage

hasDirectPart

isVariantOf

Subclass of

predecessorOf

AirCraftComp...

Sensor3

LandingSystem

Fuselage

HardwareCo...

Component

ProductModel

SpatialThing(external)

Drone,UAS, UAV

LiftGenerator...

GroundStatio...

Category2

P P P P P P P

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 15 of 19

Page 52: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

DL-Lite AC0

NC1

NP PSPACE

� Guess additional information P� Do not constructW and A explicitly� Combine T-satisfiability and R-satisfiability checks

Q0,A0 Q1,A1 . . . Qn,An Qn+1 . . .

Subclass of

Subclass of

Subclass of

Subclass of

Subclass of

hasCategory

Subclass of

isPartOf(transitive)

hasPart(transitive)

Subclass of

isLandingSystemOf

hasLandingSystemhasSensor

isSensorOf

isDirectPartOf

successorOf

Subclass of

Subclass of

hasModel

isModelOfisFuselageOf

hasFuselage

hasDirectPart

isVariantOf

Subclass of

predecessorOf

AirCraftComp...

Sensor3

LandingSystem

Fuselage

HardwareCo...

Component

ProductModel

SpatialThing(external)

Drone,UAS, UAV

LiftGenerator...

GroundStatio...

Category2

P P P P P P P

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 15 of 19

Page 53: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

DL-Lite AC0

NC1

NP PSPACE

� Guess additional information P� Do not constructW and A explicitly� Combine T-satisfiability and R-satisfiability checks

Q0,A0 Q1,A1 . . . Qn,An Qn+1 . . .

Subclass of

Subclass of

Subclass of

Subclass of

Subclass of

hasCategory

Subclass of

isPartOf(transitive)

hasPart(transitive)

Subclass of

isLandingSystemOf

hasLandingSystemhasSensor

isSensorOf

isDirectPartOf

successorOf

Subclass of

Subclass of

hasModel

isModelOfisFuselageOf

hasFuselage

hasDirectPart

isVariantOf

Subclass of

predecessorOf

AirCraftComp...

Sensor3

LandingSystem

Fuselage

HardwareCo...

Component

ProductModel

SpatialThing(external)

Drone,UAS, UAV

LiftGenerator...

GroundStatio...

Category2

P P P P P P P

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 15 of 19

Page 54: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

DL-Lite AC0

NC1

NP PSPACE

� Rewrite the R-satisfiability check for QA(i) into a first-order query QO,Pi

and answer it over the database Ai

� Use an ALOGTIME Turing machine for T-satisfiability

QO,P0

. . . . . . QO,Pn2

QO,Pn2+1 . . . . . . QO,P

n

Subclass of

Subclass of

Subclass of

Subclass of

Subclass of

hasCategory

Subclass of

isPartOf(transitive)

hasPart(transitive)

Subclass of

isLandingSystemOf

hasLandingSystemhasSensor

isSensorOf

isDirectPartOf

successorOf

Subclass of

Subclass of

hasModel

isModelOfisFuselageOf

hasFuselage

hasDirectPart

isVariantOf

Subclass of

predecessorOf

AirCraftComp...

Sensor3

LandingSystem

Fuselage

HardwareCo...

Component

ProductModel

SpatialThing(external)

Drone,UAS, UAV

LiftGenerator...

GroundStatio...

Category2

PP P

P P P P

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 16 of 19

Page 55: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

DL-Lite AC0

NC1

NP PSPACE

� Rewrite the R-satisfiability check for QA(i) into a first-order query QO,Pi

and answer it over the database Ai� Use an ALOGTIME Turing machine for T-satisfiability

QO,P0

. . . . . . QO,Pn2

QO,Pn2+1 . . . . . . QO,P

n

Subclass of

Subclass of

Subclass of

Subclass of

Subclass of

hasCategory

Subclass of

isPartOf(transitive)

hasPart(transitive)

Subclass of

isLandingSystemOf

hasLandingSystemhasSensor

isSensorOf

isDirectPartOf

successorOf

Subclass of

Subclass of

hasModel

isModelOfisFuselageOf

hasFuselage

hasDirectPart

isVariantOf

Subclass of

predecessorOf

AirCraftComp...

Sensor3

LandingSystem

Fuselage

HardwareCo...

Component

ProductModel

SpatialThing(external)

Drone,UAS, UAV

LiftGenerator...

GroundStatio...

Category2

PP P

P P P P

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 16 of 19

Page 56: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

DL-Lite AC0

NC1

NP PSPACE

� Rewrite the R-satisfiability check for QA(i) into a first-order query QO,Pi

and answer it over the database Ai� Use an ALOGTIME Turing machine for T-satisfiability

QO,P0

. . . . . . QO,Pn2

QO,Pn2+1 . . . . . . QO,P

n

Subclass of

Subclass of

Subclass of

Subclass of

Subclass of

hasCategory

Subclass of

isPartOf(transitive)

hasPart(transitive)

Subclass of

isLandingSystemOf

hasLandingSystemhasSensor

isSensorOf

isDirectPartOf

successorOf

Subclass of

Subclass of

hasModel

isModelOfisFuselageOf

hasFuselage

hasDirectPart

isVariantOf

Subclass of

predecessorOf

AirCraftComp...

Sensor3

LandingSystem

Fuselage

HardwareCo...

Component

ProductModel

SpatialThing(external)

Drone,UAS, UAV

LiftGenerator...

GroundStatio...

Category2

PP P

P P P P

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 16 of 19

Page 57: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

DL-Lite AC0

NC1

NP PSPACE

� Rewrite the R-satisfiability check for QA(i) into a first-order query QO,Pi

and answer it over the database Ai� Use an ALOGTIME Turing machine for T-satisfiability

QO,P0

. . . . . . QO,Pn2

QO,Pn2+1 . . . . . . QO,P

n

Subclass of

Subclass of

Subclass of

Subclass of

Subclass of

hasCategory

Subclass of

isPartOf(transitive)

hasPart(transitive)

Subclass of

isLandingSystemOf

hasLandingSystemhasSensor

isSensorOf

isDirectPartOf

successorOf

Subclass of

Subclass of

hasModel

isModelOfisFuselageOf

hasFuselage

hasDirectPart

isVariantOf

Subclass of

predecessorOf

AirCraftComp...

Sensor3

LandingSystem

Fuselage

HardwareCo...

Component

ProductModel

SpatialThing(external)

Drone,UAS, UAV

LiftGenerator...

GroundStatio...

Category2

PP P

P P P P

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 16 of 19

Page 58: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

DL-Lite AC0

NC1

NP PSPACE

� Rewrite the R-satisfiability check for QA(i) into a first-order query QO,Pi

and answer it over the database Ai� Use an ALOGTIME Turing machine for T-satisfiability

QO,P0

. . . . . . QO,Pn2

QO,Pn2+1 . . . . . . QO,P

n

Subclass of

Subclass of

Subclass of

Subclass of

Subclass of

hasCategory

Subclass of

isPartOf(transitive)

hasPart(transitive)

Subclass of

isLandingSystemOf

hasLandingSystemhasSensor

isSensorOf

isDirectPartOf

successorOf

Subclass of

Subclass of

hasModel

isModelOfisFuselageOf

hasFuselage

hasDirectPart

isVariantOf

Subclass of

predecessorOf

AirCraftComp...

Sensor3

LandingSystem

Fuselage

HardwareCo...

Component

ProductModel

SpatialThing(external)

Drone,UAS, UAV

LiftGenerator...

GroundStatio...

Category2

PP P

P P P P

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 16 of 19

Page 59: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

NegationDisallowing negation in TCQs:

Description

logic

Data complexity Combined complexity

0 1 2 3 0 1 2 3

DL-Lite AC0

NP

EL P NP

ALC CONP EXP

SHIQ CONP 2EXP

(Borgwardt, Lippmann, and Thost 2015; Bourgaux and Turhan 2017)

bounded history

encoding

Closed-world negation?

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 17 of 19

Page 60: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

NegationDisallowing negation in TCQs:

Description

logic

Data complexity Combined complexity

0 1 2 3 0 1 2 3

DL-Lite AC0

NP

EL P NP

ALC CONP EXP

SHIQ CONP 2EXP

(Borgwardt, Lippmann, and Thost 2015; Bourgaux and Turhan 2017)

bounded history

encoding

Closed-world negation?

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 17 of 19

Page 61: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

NegationDisallowing negation in TCQs:

Description

logic

Data complexity Combined complexity

0 1 2 3 0 1 2 3

DL-Lite AC0

NP

EL P NP

ALC CONP EXP

SHIQ CONP 2EXP

(Borgwardt, Lippmann, and Thost 2015; Bourgaux and Turhan 2017)

bounded history

encoding

Closed-world negation?

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 17 of 19

Page 62: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Metric Temporal Operators

(Baader, Borgwardt, Koopmann, Ozaki, and Thost 2017; Brandt, Kalaycı,

Kontchakov, Ryzhikov, Xiao, and Zakharyaschev 2017; Gutiérrez-Basulto, Jung, and Ozaki 2016; Thost 2018)

� Refer to specific time intervals,

e.g. “in the last 10 minutes”:

♦[−10,0] ∃f . hasLocation(d, f )∧ Hauptbahnhof(f )

� Often cause an exponential blow-up ...

� ... unless restricted to intervals [0, c], [c,∞), [−c, 0], (−∞,−c]

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 18 of 19

Page 63: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Metric Temporal Operators

(Baader, Borgwardt, Koopmann, Ozaki, and Thost 2017; Brandt, Kalaycı,

Kontchakov, Ryzhikov, Xiao, and Zakharyaschev 2017; Gutiérrez-Basulto, Jung, and Ozaki 2016; Thost 2018)

� Refer to specific time intervals,

e.g. “in the last 10 minutes”:

♦[−10,0] ∃f . hasLocation(d, f )∧ Hauptbahnhof(f )� Often cause an exponential blow-up ...

� ... unless restricted to intervals [0, c], [c,∞), [−c, 0], (−∞,−c]

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 18 of 19

Page 64: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

Metric Temporal Operators

(Baader, Borgwardt, Koopmann, Ozaki, and Thost 2017; Brandt, Kalaycı,

Kontchakov, Ryzhikov, Xiao, and Zakharyaschev 2017; Gutiérrez-Basulto, Jung, and Ozaki 2016; Thost 2018)

� Refer to specific time intervals,

e.g. “in the last 10 minutes”:

♦[−10,0] ∃f . hasLocation(d, f )∧ Hauptbahnhof(f )� Often cause an exponential blow-up ...

� ... unless restricted to intervals [0, c], [c,∞), [−c, 0], (−∞,−c]

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 18 of 19

Page 65: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

OutlookDescription

logic

Data complexity Combined complexity

0 1 2 3 0 1 2 3

DL-Lite AC0

NC1

NP PSPACE

EL P CONP NP PSPACE CONEXP

ALC CONP EXP CONEXP 2EXP

SHIQ CONP ≤EXP 2EXP

SHO[Q|I] CONP ≤ ? ≤ EXP 2EXP

ALCHOIQ CONP ≤ ? ≤ dec. CO2NEXP ≤ ? ≤ dec.SROIQ CONP ≤ ? CO2NEXP ≤ ?

Subclass of

Subclass of

Subclass of

Subclass of

Subclass of

hasCategory

Subclass of

isPartOf(transitive)

hasPart(transitive)

Subclass of

isLandingSystemOf

hasLandingSystemhasSensor

isSensorOf

isDirectPartOf

successorOf

Subclass of

Subclass of

hasModel

isModelOfisFuselageOf

hasFuselage

hasDirectPart

isVariantOf

Subclass of

predecessorOf

AirCraftComp...

Sensor3

LandingSystem

Fuselage

HardwareCo...

Component

ProductModel

SpatialThing(external)

Drone,UAS, UAV

LiftGenerator...

GroundStatio...

Category2

?

Thank you!

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 19 of 19

Page 66: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

OutlookDescription

logic

Data complexity Combined complexity

0 1 2 3 0 1 2 3

DL-Lite AC0

NC1

NP PSPACE

EL P CONP NP PSPACE CONEXP

ALC CONP EXP CONEXP 2EXP

SHIQ CONP ≤EXP 2EXP

SHO[Q|I] CONP ≤ ? ≤ EXP 2EXP

ALCHOIQ CONP ≤ ? ≤ dec. CO2NEXP ≤ ? ≤ dec.SROIQ CONP ≤ ? CO2NEXP ≤ ?

Subclass of

Subclass of

Subclass of

Subclass of

Subclass of

hasCategory

Subclass of

isPartOf(transitive)

hasPart(transitive)

Subclass of

isLandingSystemOf

hasLandingSystemhasSensor

isSensorOf

isDirectPartOf

successorOf

Subclass of

Subclass of

hasModel

isModelOfisFuselageOf

hasFuselage

hasDirectPart

isVariantOf

Subclass of

predecessorOf

AirCraftComp...

Sensor3

LandingSystem

Fuselage

HardwareCo...

Component

ProductModel

SpatialThing(external)

Drone,UAS, UAV

LiftGenerator...

GroundStatio...

Category2

?

Thank you!The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 19 of 19

Page 67: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

References IBaader, Franz, Stefan Borgwardt, Patrick Koopmann, Ana Ozaki, and Veronika Thost (2017). “Metric Temporal

Description Logics with Interval-Rigid Names”. In: Proc. FroCoS’17. Volume 10483. LNCS, pages 60–76.Baader, Franz, Stefan Borgwardt, and Marcel Lippmann (2013). “Temporalizing Ontology-Based Data Access”.

In: Proc. CADE’13. Volume 7898. LNCS, pages 330–344.— (2015a). “Temporal Conjunctive Queries in Expressive Description Logics with Transitive Roles”. In: Proc.AI’15. Volume 9457. LNAI, pages 21–33.

— (2015b). “Temporal Query Entailment in the Description Logic SHQ”. In: J. Web Semant. 33, pages 71–93.Baader, Franz, Silvio Ghilardi, and Carsten Lutz (2012). “LTL over Description Logic Axioms”. In: ACM T.Comput. Log. 13.3, 21:1–21:32.

Borgwardt, Stefan, Marcel Lippmann, and Veronika Thost (2015). “Temporalizing Rewritable Query

Languages over Knowledge Bases”. In: J. Web Semant. 33, pages 50–70.Borgwardt, Stefan and Veronika Thost (2015a). “Temporal Query Answering in DL-Lite with Negation”. In: Proc.GCAI’15. Volume 36. EPiC, pages 51–65.

— (2015b). “Temporal Query Answering in the Description Logic EL”. In: Proc. IJCAI’15, pages 2819–2825.Bourgaux, Camille and Anni-Yasmin Turhan (2017). “Temporal Query Answering in DL-Lite over Inconsistent

Data”. In: Proc. ISWC’17. Volume 10587. LNCS, pages 121–137.

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 20 of 19

Page 68: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

References IIBrandt, Sebastian, Elem Güzel Kalaycı, Roman Kontchakov, Vladislav Ryzhikov, Guohui Xiao, and

Michael Zakharyaschev (2017). “Ontology-Based Data Access with a Horn Fragment of Metric Temporal

Logic”. In: Proc. AAAI’17, pages 1070–1076.Gabbay, Dov M., Agi Kurucz, Frank Wolter, and Michael Zakharyaschev (2003). Many-Dimensional ModalLogics: Theory and Applications. Volume 148. Studies in Logic and the Foundations of Mathematics.

Gutiérrez-Basulto, Víctor, Jean Christoph Jung, and Ana Ozaki (2016). “On Metric Temporal Description

Logics”. In: Proc. ECAI’16. Volume 285. FAIA, pages 837–845.Thost, Veronika (2017). “News on Temporal Conjunctive Queries”. In: Joint Proc. of the Web Stream Processingworkshop (WSP’17) and the 2nd Int. Workshop on Ontology Modularity, Contextuality, and Evolution(WOMoCoE’17). Volume 1936. CEUR-WS, pages 1–16.

— (2018). “Metric Temporal Extensions of DL-Lite and Interval-Rigid Names”. In: Proc. KR’18. To appear.

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 21 of 19

Page 69: The Complexity Landscape of Temporal Conjunctive Queriesstefborg/Talks/HiDeSt2018_slides.pdf · High-LevelModels Abstractentities&relations Discretizedinputsignals Ontology TheComplexityLandscapeofTemporalConjunctiveQueries

PicturesErnst-Reuter-Platz by “Andreas Levers”, CC BY-NC 2.0, title slide

Taipeh Taiwan Street Head by “tingyaoh”, CC0 1.0, slides 1, 19

Drone Logistics Drone Package Drone by “geralt”, CC0 1.0, slides 1, 19

Space Center SpaceX Control Center by “SpaceX-Imagery”, CC0 1.0, slides 1, 19

Network Rectangle Rings Networking by “geralt”, CC0 1.0, slides 1, 2

Auto Sensor by “Frank Derks”, CC BY 2.0, slide 2

Artificial Neural Network Ann by “ahmedgad”, CC0 1.0, slide 2

Dronetology by “David Martín Lammerding”, CC BY 4.0, slide 4 and later

World Globe Earth Planet Blue by “qimono”, CC0 1.0, slide 17, adapted

Stopwatch Timer Clock Symbol Icon by “skeeze”, CC0 1.0, slide 18

The Complexity Landscape of Temporal Conjunctive Queries

© Stefan Borgwardt

Workshop on High-Level Declarative Stream Processing // Berlin, 25.09.2018

Slide 22 of 19