A nondeterministic semiautomaton S is predictable if there exists k ≥ 0 such that, if S knows the current input a and the next k inputs, the transition under a is deterministic. Nondeterminism may occur only when the length of the unread input is ≤ k. We develop a theory of predictable semiautomata. If a semiautomaton S with n states is k-predictable, but not (k−1)-predictable, then k ≤ (n −n)/2, and this bound can be reached for a suitable input alphabet. We characterize k-predictable semiautomata. We introduce the predictor semiautomaton, based on a look-ahead semiautomaton. The predictor is essentially deterministic and simulates a nondeterministic semiautomaton by finding the set of states reachable by a word w, if it belongs to the language L of the semiautomaton (i.e., if it defines a path from an initial state to some state), or by stopping as soon as it infers that w 6∈ L. Membership in L can be decided deterministically.
[1]
Umberto Eco,et al.
Theory of Codes
,
1976
.
[2]
Diego Calvanese,et al.
Automatic Composition of E-services That Export Their Behavior
,
2003,
ICSOC.
[3]
Derick Wood,et al.
Generalizations of 1-deterministic regular languages
,
2007,
Inf. Comput..
[4]
YO-SUB HAN,et al.
Infix-free Regular Expressions and Languages
,
2006,
Int. J. Found. Comput. Sci..
[5]
Bala Ravikumar,et al.
Deterministic Simulation of a NFA with k -Symbol Lookahead
,
2007,
SOFSEM.
[6]
Oscar H. Ibarra,et al.
Composability of Infinite-State Activity Automata
,
2004,
ISAAC.
[7]
Janusz A. Brzozowski,et al.
Predictable semiautomata
,
2009,
Theor. Comput. Sci..
[8]
M. Dal Cin,et al.
The Algebraic Theory of Automata
,
1980
.