See Table 2 (p. 683) in Robinson (1984) for values of S(p)/(p-1)! = S(p,d)/(p-1)! with p >= 2 and d = 2. In the paper, S(p) = S(p,d) is the number of (labeled) strongly connected finite automata with state set {1, 2, ..., p} and d inputs (p. 680). Since the offset here is 1, the original name of the sequence was changed to read "(n+1)-state" from "n-state".
This change agrees with
Valery A. Liskovets's formula below, who was the first one to derive expressions for the quantity S(p) = S(p,d) for a general d more than a decade before Robinson (1984). See Liskovets (1971), where S(p) = S(p,d), with d inputs, is denoted by sigma_r(n) with r = d (inputs) and n = p (number of states). For d = 2, the values of S(p) = S(p,d=2) = (p-1)!*a(p-1) for p >= 1 (with a(0) := 1) are given in
A027834, which has the correct name.
We may suggest two possible names for a(n): (i) the normalized number of labeled strongly connected (n+1)-state finite automata with 2 inputs, or (ii) the number of unlabeled strongly connected (n+1)-state finite automata with 2 inputs and a starting gate. (For purely unlabeled strongly connected n-state finite automata with 2 inputs, see
A027835, whose terms are calculated based on
Valery A. Liskovets' formulas.) (End)
