Return-Time Profiles and Quantitative Recurrence for Uniformly δ-Almost Periodic Vectors

This paper studies quantitative forms of approximate recurrence for bounded linear operators on Banach spaces through the notion of uniformly δ-almost periodic vectors. For a prescribed tolerance δ≥0, this notion relaxes classical almost periodicity by requiring uniform orbit repetitions up to an error controlled by δ, along relatively dense sets of approximate periods. The main purpose of the paper is to refine this qualitative recurrence condition by introducing return-time profiles. These profiles measure, for each accuracy level, the minimal size of recurrence windows needed to guarantee the existence of an approximate period. Thus, they provide a quantitative refinement of the usual relatively dense return condition. We prove that uniform δ-almost periodicity is equivalent to the finiteness of the associated return-time profile at every positive accuracy level. We also establish basic structural properties of these profiles, including monotonicity with respect to the accuracy and tolerance parameters, behavior under scalar multiplication and forward iteration, and an elementary additive property of approximate periods. The final part of the paper applies the general framework to weighted backward shifts on ℓp-spaces. In this setting, the explicit coordinate representation of the iterates allows us to identify several recurrence and obstruction mechanisms. We describe stable threshold recurrence, finite-support recurrence, exact recurrence generated by periodic vectors, and coordinate-level obstructions to δ-almost periodicity. The results provide a rigorous framework for measuring approximate almost periodicity in linear dynamics and clarify how recurrence-window profiles complement the classical qualitative theory of relatively dense returns.