Let's design a simplified garbage collector (GC) for a Java-like environment. We're not aiming for the full complexity of HotSpot's GC, but rather a demonstration of the underlying principles and the challenges of memory management in a managed runtime. This exercise will focus on implementing a basic Mark and Sweep garbage collector.

Imagine a simplified Java Virtual Machine (JVM) where objects are allocated on a heap. Our GC will periodically scan this heap, identify objects that are no longer reachable from the "roots" (e.g., local variables on the stack, static fields), and reclaim the memory occupied by these unreachable objects.

We'll simulate object creation and deletion. The key is to accurately track object references, handle potential cycles, and ensure thread-safety during the GC process. Our focus is on the correctness and efficiency of the garbage collection algorithm, not necessarily on performance metrics like pause times.

This is a challenging problem, requiring a good understanding of memory management, data structures, and concurrency. The real-world garbage collectors are much more complex and consider generational collection, different GC algorithms (CMS, G1), and advanced optimizations.

Requirements