Can we give the semantic domain an algebraic structure?
Let ⟦f⟧ denote the set of intervals for which M⟦f⟧(σ) = tt, i.e.,
⟦f⟧ ≜ {σ | M⟦f⟧(σ) = tt}
The ∨ of two PITL formula is then
| ⟦f1 ∨ f2⟧ =                                           | 
| use  definition of ⟦ ⟧               | 
| {σ  | M⟦f1 ∨ f2⟧(σ) = tt}           | 
| use  definition of M⟦f1 ∨ f2⟧(σ)                 | 
| {σ  | M⟦f1⟧(σ) = tt or M⟦f2⟧(σ) = tt}    | 
| use  set theory, let  ∪ denote union             | 
| {σ  | M⟦f1⟧(σ) = tt}∪{σ  | M⟦f2⟧(σ) = tt} | 
| use  definition of ⟦ ⟧               | 
| ⟦f1⟧∪⟦f2⟧                    | 
So we need algebraic operators that correspond to ¬, ∨,skip,; and ∗
¬ corresponds to complement of a set, i.e.,
| ⟦¬f⟧ =                                                     | 
| use  definition of ⟦ ⟧                | 
| {σ  | M⟦¬f⟧(σ) = tt}                | 
| use  definition of M⟦¬f⟧(σ)                           | 
| {σ  | not (M⟦f⟧(σ) = tt)}             | 
| use  set theory, let ∼A denote set complement | 
| ∼{σ  | M⟦f⟧(σ) = tt}               | 
| use  definition of ⟦ ⟧                | 
| ∼⟦f⟧                         | 
What about chop (‘;’)?
Let ⋅ denote the fusion of two intervals σ1,σ2 ∈ Σ+ ∪ Σω, i.e.,
Let a,b ∈ Σ (a and b are not the same), v,w ∈ Σ∗ and s,t ∈ Σω
σ1 ⋅σ2 ≜
| vaw | if  | σ1 = va, | σ2 = aw | 
| ∅   | if  | σ1 = va, | σ2 = bw | 
| vas   | if  | σ1 = va, | σ2 = as  | 
| ∅   | if  | σ1 = va, | σ2 = bs  | 
| s     | if  | σ1 = s,  | σ2 = aw | 
| s     | if  | σ1 = s,  | σ2 = t   | 
Let S,T ⊆ Σ+ ∪ Σω then
S ⋅T ≜{σ1 ⋅ σ2 | σ1 ∈S and σ2 ∈T}
‘;’ corresponds to fusion ‘⋅’, i.e.,
| ⟦f1 ; f2⟧ =                                                                        | 
| use  definition of ⟦ ⟧                           | 
| {σ  | M⟦f1 ; f2⟧(σ) = tt}                        | 
| use  definition of M⟦f1 ; f2⟧(σ)                                              | 
| {σ  | ( exists k, s.t. M⟦f1⟧(σ0…σk) = tt and M⟦f2⟧(σk…σ|σ|) = tt) | 
|    or (σ is infinite  and M⟦f1⟧(σ) = tt)}               | 
| use  definition of ⋅                            | 
| {σ  | M⟦f1⟧(σ) = tt}⋅{σ  | M⟦f2⟧(σ) = tt}             | 
| use  definition of ⟦ ⟧                           | 
| ⟦f1⟧⋅⟦f2⟧                                 | 
What about empty?
| ⟦empty⟧ =                           | 
| use  definition of ⟦ ⟧       | 
| {σ  | M⟦empty⟧(σ) = tt}    | 
| use  definition of M⟦empty⟧(σ) | 
| {σ  | σ is a 1 state interval}  | 
| use  definition of Σ                | 
| Σ  | 
What about skip?
skip can be defined as ∼(Σ ∪∼Σ ⋅∼Σ)
| ∼(Σ ∪∼Σ ⋅∼Σ) =              | 
| use  De Morgan for set theory | 
| ∼Σ ∩∼(∼Σ ⋅∼Σ)              | 
∼Σ is the set of intervals containing ≥ 2 states
∼Σ ⋅∼Σ is the set of intervals containing ≥ 3 states
∼(∼Σ ⋅∼Σ) is the set of intervals containing ≤ 2 states
∼Σ ∩∼(∼Σ ⋅∼Σ) is the set of intervals containing exactly 2 states
What about a state formula, i.e., a formula without temporal operators?
A state formula only constrains the first state of an interval. Let p be a state formula. Then the following holds
⟦p⟧ = (⟦p⟧∩Σ) ⋅T where
| T ≜⟦true⟧ = Σ+ ∪ Σω | 
| ∅≜⟦false⟧ = ∅      | 
What about chopstar ‘∗’?
In the semantics of ‘∗’ both finite and infinite iteration are considered simultaneously. Let’s define separate algebraic operators for them.
Let S∗ and Sω denote respectively finite and infinite iteration of a set S ⊆ Σ+ ∪ Σω and can be defined as follows
| finite | ≜ | ∼(T ⋅∅)               | fmore | ≜ | ∼Σ ∩finite      | 
| f(X) | ≜ | Σ ∪ (S ∩fmore) ⋅X | g(X) | ≜ | (S ∩fmore) ⋅X | 
| S∗   | ≜ | μX.f(X)              | Sω    | ≜ | νX.g(X)         | 
where μX.f(X) denotes the least fixpoint of f and νX.g(X) the greatest fixpoint of g.
Then we have
⟦f∗⟧ = … = ⟦f⟧∗∪⟦f⟧ω