In per-survivor processing (PSP), the number of parameter estimators e
quals the number of retained hypothetical data sequences (survivors).
In this paper, we propose an algorithm which uses an arbitrary number
of parameter estimators and compromises between the two extremes of te
ntative decisions (one parameter estimator) and PSP. Specific applicat
ions to reduced state sequence estimation (RSSE) and maximum likelihoo
d sequence estimation (MLSE) with adaptive tracking of a fading channe
l are considered.