Add simple binary utility for contact-rate policy tradeoff

[gvegayon]

Motivation

• Expose a concrete, optimization-ready utility that encodes the tradeoff between economic activity (contacts) and transmission risk for binary group policies.
• Replace the placeholder utility with a simple, solver-agnostic evaluator that can be used by downstream optimizers or search routines.

Description

• Add tunable weights economic_weight and health_weight to ABM and implement ABM::evaluate_policy_utility(const std::vector<int>&) to evaluate binary policies (0 = no reduction, 1 = reduced contacts).
• Implement utility as economic_weight * average_contact_multiplier - health_weight * (p_t * c_eff * i * (1-i)), where c_eff is the effective contact rate and i is infected share, and validate policy length/entries with std::invalid_argument.
• Fix Group::sample_infected() contact-rate application so reduced-contact groups use the reduction multiplier, update m_calc_utility() to construct current policy and call evaluate_policy_utility(), and add necessary headers (<numeric>, <stdexcept>).
• Add documentation artifacts: a ready-to-post issue draft (docs/utility-function-issue.md), a proposal with literature context (docs/utility-function-proposal.md), and a short README note pointing to ABM::evaluate_policy_utility.

Testing

• Built the project with make main.o, which completed successfully.
• Ran the model with ./main.o and confirmed it executed and produced the epidemic curve output (run completed successfully).
• Created the in-repo issue draft because the GitHub CLI is not available in the environment, so the issue content is in docs/utility-function-issue.md (manual posting required).

---

Codex Task