Skip to content

solutions

The solutions module provides the full infrastructure for solving cadet–AFSC matching problems in the AFCCP system. It includes a range of algorithms (HR, GA, GUO, etc.), constraint handling logic, Pyomo-based model construction and solver interfaces, and extensive support for sensitivity analysis and "what-if" scenarios.

This module is central to the AFCCP ecosystem—it orchestrates the end-to-end optimization workflow: generating initial solutions, building and solving models, evaluating results, and exploring trade-offs through iterative constraints or shifting objectives.

Submodules

  • solutions.algorithms Implements baseline and advanced matching algorithms including:
    • Classic and enhanced Hospital/Residents (HR) matching
    • GA/GMA: Genetic algorithms with cadet–AFSC crossover, mutation, and selection logic
    • Hybrid solutions: Merge utility-based models with HR/GMA results for feasible fallback options
  • solutions.handling Provides tools for:
    • Evaluating objective values: VFT scores, fairness metrics, quota and eligibility violations
    • Constraint diagnostics: Identifies which constraints were violated or binding
    • Preference reconstruction: Converts j_array solutions into readable cadet assignments
  • solutions.optimization Contains the core Pyomo model logic and solver interfaces for:
    • VFT models (exact and approximate)
    • GUO (Global Utility Optimization) and goal programming models
    • Constraint injection, warm-starting, and solution persistence
    • Full compatibility with real-world constraints like AFOCD tiers, base/training thresholds, and USAFA rules
  • solutions.sensitivity Tools for constraint iteration, GA initialization, and what-if analysis, including:
    • Automated constraint relaxation workflows
    • PGL capacity adjustment loops and visualizations
    • "What If List.csv" batch simulation and delta metric reporting

Typical Workflow

  1. Generate Initial Solution(s)
    • Use algorithms to run HR or GA on initial cadet–AFSC preferences
    • Or pre-seed GA populations using VFT/Assignment heuristics from sensitivity
  2. Build and Solve Optimization Model
    • Construct VFT or GUO model in optimization
    • Add constraints based on cadet thresholds, base/training requirements, or fairness tiers
  3. Evaluate Solution
    • Use handling to check feasibility, score each objective, and extract readable results
  4. Run Sensitivity Analysis
    • Apply sensitivity to modify constraints, quotas, or preference structures
    • Generate what-if metrics and comparative charts

See Also

  • data.preferences: Cadet/AFSC eligibility, tier data, and OM-based utility scores

  • data.values: Solver weight configurations for VFT and GUO objectives

  • CadetCareerProblem: Main interface to run all solution workflows via .solve_*() methods