OQ-mef.8: add combinatorial optimization adapters (AFML Ch.21)

[Sean-Koval]

Summary

• add combinatorial_optimization module for AFML Chapter 21 discrete problem support
• implement integer-encoded decision schema, pluggable objective trait, exact finite-set solver, and solver-adapter interface
• add trading-trajectory state-space enumeration + exact solver + dynamic path-dependent objective evaluation with fixed ticket costs (non-convex)
• add integration tests covering exact search, adapter gap, trajectory objective behavior, and path-limit safeguards
• mark Chapter 21 as done in docs metadata and add module docs (formulas, dynamic trajectory example, scalability caveats)

Validation

• cargo test -p openquant --test combinatorial_optimization
• cargo test -p openquant --lib --tests
• PYO3_USE_ABI3_FORWARD_COMPATIBILITY=1 cargo clippy --all-targets --all-features -- -D clippy::correctness -D clippy::suspicious