perf: optimize StateSet.equals to eliminate heap allocations during a…

[rajat315315]

Closes: #16413

📊 Benchmark & GC Allocation Metrics

Evaluated using ThreadMXBean.getThreadAllocatedBytes() and GarbageCollectorMXBean across 30 determinization runs on complex NFA union automata (1,000 regex terms with overlapping state sets):

GC / Memory Metric	Baseline (Before Optimization)	Optimized (Option C)	Improvement / Reduction
Total Heap Memory Allocated (30 runs)	548.72 MB	100.42 MB	81.7% reduction (~448.3 MB saved!)
Avg Heap Allocated per determinize()	18.29 MB (19,179,268 bytes)	3.35 MB (3,510,065 bytes)	Saved ~14.94 MB of garbage per call!
GC Collection Cycles	11 GC cycles	2 GC cycles	81.8% reduction in GC pauses
Total GC Pause Time	142 ms	25 ms	82.4% reduction in GC pause duration
Average Execution Time	140.43 ms	76.30 ms	~1.84x faster execution speed

[rajat315315]

Attaching the Benchmark file used to generate the results.
TestDeterminizeBenchmark.java