"Correct" formulation of rules is relative to some context, for example in the context of a specific theory like Distributed Morphology, or Two Level Morphology. If you are just attempting to describe a fact pattern, a plain English statement would suffice, and in fact is superior to a "formalization", since using a set of symbols the way you did suggests that you are operating in some formal theoretical framework, which doesn't seem to be the case.
The usual way of treating the pattern that you describe is to affix -i in the morphology (no further conditions), and have a separate phonological rule of schwa-deletion (before a vowel, possibly just /i/). That is, "affix -i" regardless of context, then ə→Ø/__V. You don't say what happens with vowel-final words, and that makes a difference. In certain versions of Distributed Morphology, there could not be a rule that simply changes final schwa to i, there has to be a specific triggering context, which would probably lead to positing a general affix /i/ with a peculiar phonological property (in case there are other invariant affixes of the shape /i/ in the language). Work by Shanti Ulfsbjörninn has pushed this particular line of reasoning. In other words, correctness depends on your theoretical assumptions.
