Content deleted Content added
No edit summary |
m Dating maintenance tags: {{Clarify span}} {{Clarify}} |
||
(8 intermediate revisions by 3 users not shown) | |||
Line 2:
== Syntax ==
The '''future fragment of TPTL''' is defined similarly to [[linear temporal logic]], in which furthermore, [[clock (model checking)|clock]] variables can be introduced and compared to constants. Formally, given a set <math>X</math> of clocks, {{clarify span|MTL|reason=Should be 'TPTL'? MTL is discussed in a later section.|date=February 2025}} is built up from:
* a finite set of [[propositional variable]]s ''AP'',
* the [[logical connective|logical operators]] ¬ and ∨, and
* the [[Temporal logic|temporal]] [[modal operator]] '''U''',
* a clock comparison <math>x\sim c</math>, with <math>x\in X</math>, <math>c</math> a number and <math>\sim</math>
* a freeze quantification operator <math>x.\phi</math>, for <math>\phi</math> a TPTL formula with set of clocks <math>X\cup\{x\}</math>.
Furthermore, for <math>I=(a,b)</math> an interval, <math>x\in I</math> is considered as an abbreviation for <math>x>a\land x<b</math>; and similarly for every other kind of intervals.
The logic '''TPTL+Past'''<ref name="ExpressivenesOfTPTLAndMtl">{{cite
* the temporal modal operator '''S'''.
A '''closed formula''' is a formula over an empty set of clocks.<ref name="Really">{{cite journal |last1=Alur |first1=Rajeev |author1link = Rajeev Alur|last2=Henzinger |first2=Thomas A. |author2link = Thomas Henzinger|title=A really temporal logic|journal=[[Journal of the ACM]]|date=Jan 1994 |volume=41 |issue =1|pages=181–203 |doi=10.1145/174644.174651|doi-access=free }}</ref>{{clarify|reason=The set of clocks of a formula needs to be introduced (I guess along with the syntax case distinction above).|date=February 2025}}
==
Let <math>T\subseteq\mathbb R_+</math>,
==
Let <math>T</math> and <math>\gamma</math> be as above. Let <math>X</math> be a set of [[clock (model checking)|clocks]]. Let <math>\nu:X\to\mathbb R_{\ge0}</math> (a ''clock valuation'' over <math>X</math>).
We are now going to explain what it means
Let <math>\phi</math> and <math>\psi</math> be two formulas over the set of clocks <math>X</math>, <math>\xi</math> a formula over the set of clocks <math>X\cup\{y\}</math>, <math>x\in X</math>, <math>l\in\mathtt{AP}</math>, <math>c</math> a number and <math>\sim</math> being a comparison operator such as <, ≤, =, ≥ or >:
We first consider formulas whose main operator also belongs to LTL:
* <math>\gamma,t,\nu\models l</math> holds if <math>l\in\gamma(t)</math>,
* <math>\gamma,t,\nu\models\neg\phi</math> holds if
* <math>\gamma,t,\nu\models\phi\lor\psi</math> holds if either <math>\gamma,t,\nu\models\phi</math> or <math>\gamma,t,\nu\models\psi</math>, or both
* <math>\gamma,t,\nu\models\phi\mathbin\mathcal U\psi</math> holds if there exists <math>
* <math>\gamma,t,\nu\models\phi\mathbin\mathcal S\psi</math> holds if there exists <math>t''
* <math>\gamma,t,\nu\models
* <math>\gamma,t,\nu\models
== Metric temporal logic ==
[[Metric temporal logic]] is another extension of LTL
TPTL is as least as expressive as MTL. Indeed, the MTL formula <math>\phi\mathbin\mathcal {U_I}\psi</math> is equivalent to the TPTL formula <math>x.\phi\mathcal(x\in I\land\psi)</math> where <math>x</math> is a new variable.<ref name="Really"/>
It follows that any other operator introduced in the page [[metric temporal logic#Operators defined from MTL operator|MTL]], such as <math>\Box</math> and <math>\Diamond</math> can also be defined as TPTL formulas.
|