On the Outcome Map of MVP Parking Functions: Permutations Avoiding 321 and 3412, and Motzkin Paths

Abstract

We introduce a new parking procedure called MVP parking in which n cars sequentially enter a one-way street with a preferred parking spot from the n parking spots on the street. If their preferred spot is empty they park there, otherwise a later car bumps a previous cars out of their preferred spot sending that car forward in the street to find the first available spot on the street. If all cars can park under this parking procedure, we say the list of preferences of the n cars is an MVP parking function of length n. We show that the set of (classical) parking functions is exactly the set of MVP parking functions although the parking outcome (order in which the cars park) is different under each parking process. Motivating the question: Given a permutation describing the outcome of the MPV parking process, what is the number of MVP parking functions resulting in that given outcome? Our main result establishes a bound for this count which is tight precisely when the permutation describing the parking outcome avoids the patterns 321 and 3412. We then consider special cases of permutations and give closed formulas for the number of MVP parking functions with those outcomes. In particular, we show that the number of MVP parking functions which park in reverse order (that is the permutation describing the outcome is the longest word in Sn, which does not avoid the pattern 321) is given by the nth Motzkin number. We also give families of permutations describing the parking outcome for which the cardinality of the set of cars parking in that order is exponential and others in which it is linear.