Back

Honours | Calendar | References (Normal) (Formal) (Quick) (Links only) (Light Toggle)

Total of 31 records found.
Current page is: file:///C:/My Documents/Education/University/public_html/comp471.html?HTML=true&
File is: comp471_lite.html

Sort by Category, type, title, short_title, author, pub_date, read_date, rank. View rank.


Paper[LN] Title: COMP 471: Semantics of Concurrency - Lecture Notes
Author(s): Matthew Hennessy
Publication Date: (2002)
Remote: http://www.mcs.vuw.ac.nz/courses/COMP471/2002/
Comments: Lecture notes.


term[BISIM] Title: Bisimulation
Comments: To prove a machine X simulates Y, come up with a relation between states in X and Y and then show it works.
To prove otherwise, show that no possible relation will have a simulation that works.
X should simulate Y and Y should simulate X. The relations should be the inverse of each other.


term[LTS] Title: Labelled Transition Systems (LTS)
Found/Read: 2002/02/26
Comments: From the lecture notes [LN].
An lts is a triple (Q, Act, T).
A formalism to encompass all the different views of a system.


term[LEQ] Title: Language equivalence
Comments: Two states are language equivalent if they can generate exactly the same set of strings.
Language equivalenve is fine for finite automata, but not so suitable for more general systems. Example of behavioural equivalence of two drink machines.



term[SIM] Title: Simulation
Comments: From [LN]:
Definition 1 In a given LTS [LTS] the state Q simulates the state P if there is a simulation (relation) which relates P and Q; that is a simulation S such that P S Q.
For Q to simulate P we require:

To prove that S is a simulation we need to show that:

there is some corresponding action from the second state B -a-> B' such that the pair (A',B') is again in S.

To show that a certain state does not simulate another state. Show that Q cannot simulate all the moves that P can. I.e. P can do the same action to two different states where as Q can only do the action two one state. The important part comes from P having more options than Q.
A dead state can be simulated by almost any state, as by definition it can't do any actions.
See also [SBISIM]


term[SBISIM] Title: Strong Bisimulation Equivalence
Comments: P ~ Q
A notion of behavioural equivalence between system states should be symmetric. Strong bisimulation is just a symmetric version of simulation [SIM].

R is a simulation. R-1 is also a simulation, R-1 = {(Q,P) | (P,Q) ∈ R}
Definition 2 In an lts [LTS] we say the states P and Q are strongly bisimular, written P ~ Q, if there is some bisimulation R such that P R Q.

If P simulates Q and Q simulates P it is not necessarily true that P and Q are bisimilar, the relation in one direction may no be the reverse of the other.


term[WBE] Title: Weak Bisimulation Equivalence
Comments: a.k.a. Observational Equivalence. P ~~ Q. Ignoring internal moves (τ) - treats them as unobservable.
P =s=> P'
P =ε=> P' ≡ P -τ*-> P'
P =a1…an=> Q ≡ P -τ*-> -a1-> -τ*-> … -τ*-> -an-> -τ*-> Q

P =û=> P' ≡ P -τ-> P' if μ = τ
P =û=> P' ≡ P =μ=> P' if μ ≠ τ
<<μ>>Φ - Satisfied by a process P if P =û=> P' for some residual P' such that P' statisfies Φ.


term[LPT] Title: Language of process terms - sequential processes
Comments: P ::= 0 | µ.P | P + P | A(ñ)
µ ::= n? | n! | n
A(ñ) <= DA


term[SEM] Title: Semaphore
Comments: A Historical Note
Semaphores were first proposed by Edsger W. Dijkstra in 1968. Dijkstra, a Dutchman, used P for Wait() because P is the first letter of the Dutch word "passeren," which means "to pass," and V for Signal() because V is the first letter of the Dutch word "vrijgeven," which means "to release." Some books may also use down and up for Wait() and Signal(), respectively.


term[CSP] Title: The Critical Section Problem
Comments: Mutual exclusion - only one process at a time in a critical section
Progress - some process must always be making process
Bounded waiting - no process should be delayed indefinitely.


term[ML] Title: Modal Logic
Comments: Φ ::= true | ¬Φ | Φ ∧ Φ | <μ>Φ
false = ¬true
P |= <μ>Φ' if P -μ-> P' where (there exists a P' such that) P' |= &Phi' -it is possible to do μ and enter a state where Φ is true.
[a]Φ = ¬<a>(¬Φ) - everytime an 'a' is done &Phi is true.
The formula [μ]Φ is satisfied by by a process if every state it can reach satisfies the property Φ. [a]false is satisfied by a process if it can not do the action a.

See also [EML]


term[LCS] Title: A Language for Concurrent Systems
Comments: A language for discribing processes running in parallel.
P ::= 0 | μP | P + P | P|P | (new c) P
Syncronisation (τ - the syncronisation of two processes on a local resource)(Communication) between subprocesses is achieved via complmentary (input/output) pairs.


term[LE] Title: Logical Equivalence
Comments: Slide 21 [LN] - L(p) = { Φ | p |= Φ}


term[MUC] Title: μ-calculus
Comments:

See also [FPT]


term[PO] Title: Partial Order
Comments: A partial order consists of a pair (P, ≤), where P is a set and ≤ is a relation over P which satisfies:
Reflexivity: p ≤ p
Anti-symmetry: p1 ≤ p2 and p2 ≤ p1 then p1 = p2
Transitivity p1 ≤ p2 and p2 ≤ p3 implies p1 ≤ p3

Powerset - P(Q) = {A | A ⊆ Q} - the set of all subsets of Q
We say p ∈ P is an upper bound for S if q ≤ p for every element q of S. The least upper bound of S is the smallest upper bound and is denoted ∨S.
We say p ∈ P is an lower bound for S if p ≤ q for every element q of S. The greatest lower bound of S is the largest lower bound and is denoted ∧S - p' ≤ p for every lower bound p' of S.

See also [MONO] [CL]


term[CL] Title: Complete lattice
Comments: Defintion 1 (Complete lattice) The partial order [PO] (P, ≤) is a complete lattice if every subset of P has a least upper bound.
(P(Q),⊆) is a complete lattice. = Q ⊥ =
(P,≤) has a largest element which is its least upper bound, ∨P and is denoted . Least Upper Bound ∨S={A | A ∈ S} - the union of all the sets in S.
(P,≤) has a least element which is its greatest lower bound, ∧P and is denoted ⊥. Greatest Lower Bound S=∩{A | A ∈ S} - the set intersection.
Proposition 2 In a complete lattice (P, ≤) every subset of P has a greatest lower bound.


term[MONO] Title: Monotonic
Comments: Definition 3 A function ƒ : P → P, where P is a partial order, is called monotonic if ƒ(p1) ≤ ƒ(p2) whenever p1 ≤ p2.


term[FPT] Title: Fixpoint Theorems
Comments: [comp462 FOLDOC remote] "<mathematics> The fixed point of a function, f is any value, x for which f x = x. A function may have any number of fixed points from none (e.g. f x = x+1) to infinitely many (e.g. f x = x). The fixed point combinator, written as either "fix" or "Y" will return the fixed point of a function."

Definition 4 (Fixpoints) Let ƒ be any function over a complete lattice [CL] (P,≤). We say the element p of P is a fixpoint of ƒ if ƒ(p) = p
Maximal fixpoints dominate all others. That is it is a fixpoint and p' ≤ p for ever fixpoint p' of ƒ.
Minimal fixpoints dominate all others. That is it is a fixpoint and p ≤ p' for ever fixpoint p' of ƒ.
Theorem 5 Every monotonic function over a complete lattice has both a maximal and a minimal fixpoint.

See also [PO]


term[CC] Title: co-chain
Comments: A co-chain is a sequence of points in a complete lattice [CL] of the form p0 ≥ p1 ≥ p2 ≥ p3 ≥ …
A co-chain may be finite or infinte.
Definition 6 (co-continuous functions) A function ƒ over a complete lattice [CL] P is called co-continuous if for every non-empty co-chain C of P ∧{ƒ(q) | q ∈ C } = ƒ(∧C)

pω = ∧{pn | n ∈ ω} p0 =
Theorem 7 pω is the maximal fixpoint of the co-continuous function ƒ.


term[EML] Title: Extended Modal Logic
Comments: Φ ::= true | false | X ∈ Vars | Φ ∨ Φ | Φ ∧ Φ | <K>Φ | [K]Φ | (minX)Φ | (maxX)Φ
M[[Φ]] - the set of all states which satisfy Φ - if q ∈ M[[Φ]] then q |= Φ
¬<a>Φ → [a]¬Φ
¬[a]Φ → <a>¬Φ
<-> = <Act> - all possible actions.
[-] = [Act]
M[[<->true]] = all states that can do at least one action (non-dead states)
M[[ [-]false ]] = set of all dead states.
M[[ [-]Φ ∧ <->true]] = Set of states that are alive and after any action they can do the reached state can satisfy Φ
M[[ [-]X ]](S) = The set of states that for any action they can perform, will end up in a state in S. This will include all dead states.

M[[ <K>Φ ]] = {q ∈ Q | q -a-> q', for some a ∈ K and q' ∈ M[[ Φ ]](S)} - this is true for states where there exists one action from K that leads to a state that satisfies Φ
M[[ [K]Φ ]] = {q ∈ Q | q -a-> q', for any a ∈ K, implies q' ∈ M[[ Φ ]](S)} - this is true for states where for every action in K, they can reach a state that satisfies Φ. If the state can't do any actions in K (like a dead state) then this is also true.

See also [ML]


term[MAXF] Title: Maximal Fixpoint
Comments: The first approximation to a maximal fixpoint is always [CL].

Example:
M[[ A ∧ [-]X ]] - The set of all states that can perform the action a and all the successors of these states can also perform the action a. <a>true is an invarient of these states, saftey properties.
importantT(Φ) = (maxX) (Φ ∧ [-]X) - Then p |= T(Φ) if and only if p' |= Φ for every p' such that p -Act-> p'.
M[[ A ]](S) ∩ M[[ [-]X ]](S)
p0 = = All the States in the LTS [LTS].
pn = M[[ A ]](S) ∩ M[[ [-]X ]](pn-1)
Until pn=pn-1 - Stable point pω - maximal fixpoint is reached.

See also [SAFE]


term[MINF] Title: Minimal Fixpoint
Comments: The first approximation to a minimal fixpoint is always [CL].

Example:
M[[ D ∨ <X> ]] - The set of all states that can perform the action d or has a successor that can perform the action d.
importantT(Φ) = (minX)(Φ ∨ <->X) - Then q |= T(Φ) if and only if q -Act->* q' for some q' such that q' |= Φ.
p0 = =
pn = M[[ D ]](S) ∪ M[[ <->X ]](pn-1)
Until pn=pn-1 - Stable point pω - minimal fixpoint is reached.

M[[ [-]X ]]() = - pg 10 of mu-calculus notes.

See also [LIVE]


term[LIVE] Title: Liveness properties
Comments: Some specific property will eventually become true. Minimal fixpoints [MINF].
(minX)(Dead ∨ <->X) - true of states that can lead to deadlock via some sequence of actions.
(minX)(Φ ∨ (<->true ∧ [-]X)) - the property Φ will eventually be true on all paths.


term[SAFE] Title: Safety properties
Comments: Some specific property is true and will remain true. An invarient of some states. Maximal fixpoints [MAXF].
(maxX) (Φ ∧ [K]X) - the property Φ remains true so long as only acitons from the set K are performed.
(maxX) (Φ' ∨ (Φ ∧ [-]X)) - the property Φ will continue to hold regardless of what actions are performed, so long as Φ' remains false.


term[CCS] Title: Calculus for Communicating Logics
Publisher: Edinborgh.


term[VCCS] Title: Value Passing CCS
Found/Read: 2002/05/21
Comments: Not examinable...
P ::= 0 | c?x.P | c!e.P | P + P | P|P | (new c) P | if e1 = e2 then P else Q | A(ñ,e)
When handshaking occurs a value is passed over. Can be a tuple of values (including local channels for privacy).
Online sorting and RPC examples


term[LAWSR] Title: Laws for Recursively defined processes
Found/Read: 2002/05/16
Comments: Not examinable...
Unwinding substitue in the definition.

τ-laws μ.τ.P = μ.P
P + τ.P = τ.P
μ(P + τ.Q) = μ.(P + τ.Q) + μ.Q
μ may be τ or an external action.

Monoid Laws P + Q = Q + P
(P + Q) + R = P + (Q + R)
P + P = P
P + 0 = P

Interleaving For (new c)(P|Q) - take term outside if it is not boudn locally. If it is, only take it out when complimentary pairs exist as the next actions in two processes.


term[IP] Title: infinte path
Comments: [α]Φ - this is true of any state that can't do a α
If α is Act (all possible actions in the LTS) then [α]Φ is true of any dead states.
[α] is true of any state that can't perfrom any of the actions in α.
So (minX) (A ∨ [b]X) will be true of any state where via a sequence of b's it is possible to da an a or a state can be reached where it is not possible to do a b - hence creating a infinite path.


term[PLOGIC] Title: Property logics


term[ATHEORY] Title: Algebraic theories


term[PI] Title: The picalculus
Comments: Not examinable...


^Top^

Honours | Calendar | References (Normal) (Formal) (Quick) (Links only) (Light Toggle)