This is ultimately an optimization issue. You might not want to handle a huge open list storage space if 99.9% of your searches will only use 1% of that storage space (so the remaining 0.1% of your searches would allocate some memory to work correctly, but is this so frequent that it becomes problematic?)

I'd measure what open list size is encountered by your program when it's running, then choose a storage space size that can support 99.9% of these searches, and use a dynamic storage space when the bounds are exceeded.