Download presentation
Presentation is loading. Please wait.
Published byGiles Day Modified over 9 years ago
1
Automated Reasoning SS08 Christoph Weidenbach TexPoint fonts used in EMF. Read the TexPoint manual before you delete this box.: A A A
2
Christoph Weidenbach Automated Reasoning SS08 Content Propositional LogicDPLL2-Watch, Learning Propositional Logic + Theories DPLL(T)Coupling First-Order LogicSuperposition Indexing, Sharing, Filtering First-Order Logic + Theories SUP(T)Coupling Linear Arithmetic LogicCalculusAlgorithms
3
Christoph Weidenbach Automated Reasoning SS08 Propositional Logic & & 1 1 P Q R
4
Christoph Weidenbach Automated Reasoning SS08 Hardware Industrial Processor Verification: 14-cycle Model Checking 1Mio Variables, 10 Mio Literals, 4 Mio Clauses 3 hours run time (2004)
5
Christoph Weidenbach Automated Reasoning SS08 SUDOKU 1 2 3 4 5 6 7 8 9 123456789123456789
6
Christoph Weidenbach Automated Reasoning SS08 Summary propositional logic is suitable to represent finite domains software: restrict all variables to finite domain hardware: restrict number of cycles suitable to test problems with thousands of variables Limits: infinite domains or “calculations”, i.e., mathematical structure
7
Christoph Weidenbach Automated Reasoning SS08 Propositional Logic + T
8
Christoph Weidenbach Automated Reasoning SS08 Dutch Soccer League if Eindhoven and Amsterdam play on the same day the TV income is x If Eindhoven and Amsterdam play on two different days, the income is 2x if a team plays on Wednesday champions league it doesn’t play on Friday there are at most 3 plays on Friday ….. in sum several thousand constraints over LP and Boolean variables League is modelled by the Barcelogic tool
9
Christoph Weidenbach Automated Reasoning SS08 Transition Systems
10
Christoph Weidenbach Automated Reasoning SS08 Summary propositional logic + T can also represent aspects of infinite theories for the “meta algorithm” for the theory often “nice” properties are needed bottleneck often the solver for T Limits: quantification and “free” structures beyond boolean combinations
11
Christoph Weidenbach Automated Reasoning SS08 First-Order Logic All professors love squeezing all students. Chris is a professor.
12
Christoph Weidenbach Automated Reasoning SS08 SUDOKU 2, 3, 4 9595 1, 4, 6, 8 7
13
Christoph Weidenbach Automated Reasoning SS08 SUDOKU
14
Christoph Weidenbach Automated Reasoning SS08 SUDOKU 7 1 2 3 4 5 6 7 8 9 123456789123456789
15
Christoph Weidenbach Automated Reasoning SS08 SUDOKU 7 1 2 3 4 5 6 7 8 9 123456789123456789
16
Christoph Weidenbach Automated Reasoning SS08 LAN Router Sent(epacket(incoming-net, router-mac, src-mac, e-ip, ippacket(ip-src, ip-dst, ip-proto, ip-data)))) RouteEntry(route(router,dst-netmask,dst-net-addr,outgoing-net)) ipand(ip-dst,dst-netmask) dst-net-addr Sent(epacket(outgoing-net, dst-mac, src-mac, e-ip, ippacket(ip-src, ip-dst, ip-proto, ip-data)))
17
Christoph Weidenbach Automated Reasoning SS08 Summary first-order logic can model freely defined infinite theories inductive theories are out of scope incredible expressiveness full quantification Limits: undecidable (take serious), some important theories can not be (finitely) represented
18
Christoph Weidenbach Automated Reasoning SS08 First-Order Logic + T All professors above 50 love squeezing all students. Chris is a professor below 50.
19
Christoph Weidenbach Automated Reasoning SS08 Hybrid Automata
20
Christoph Weidenbach Automated Reasoning SS08 Summary first-order logic + T can also model aspects involving inductive theories adequately represent many aspects of software, hardware incredible incredible expressiveness quantification over theory variables potentially limited Limits: very undecidable (take serious), in general not compact, and any calculus not complete
21
Christoph Weidenbach Automated Reasoning SS08 The End: Let’s Start Propositional LogicDPLL2-Watch, Learning Propositional Logic + Theories DPLL(T)Coupling First-Order LogicSuperposition Indexing, Sharing, Filtering First-Order Logic + Theories SUP(T)Coupling Linear Arithmetic LogicCalculusAlgorithms
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.