Add Max-Cut optimization task

[Wu-kong-123]

This PR adds a new optimization task: Max-Cut.

[Wu-kong-123]

Max-Cut Problem: Given an undirected graph $G = (V, E)$ with non-negative weights $w_{ij}$ assigned to each edge $(i, j) \in E$, the objective is to partition the set of vertices $V$ into two disjoint subsets, $S$ and $T$ (where $T = V \cup S$), such that the total weight of the edges connecting the two sets is maximized.

---

Experimental Evaluation

Experimental Setup:

• LLM: deepseek-v4-flash
• Problem Scale: n_instance = 16, n_node = 100
• Time Constraints: timeout_second = 60
• Hyper-parameters:
  • Common: max_sample_nums = 100, pop_size = 5

Results:

Experimental results confirm that the Max-Cut task is compatible with existing frameworks (EoH and MCTS-AHD) and runs successfully, validating its readiness for use in LLM4AD.