How changxincao/tsppsvu works
TSPP-SVU is research software for transportation-service procurement under contextual stochastic uncertainty. Its core workflow prepares historical or generated instances, represents demand scenarios and procurement parameters, solves a procurement optimization model, and records decisions and diagnostics. The Java model layer spans SAA/CSAA, RCSAA, divergence-DRO, Wasserstein-DRO, and decomposition approaches; Python helpers support random-forest weighting and partial cross-moment experiments. The README documents formal experiment entry points, but those runners are absent from the supplied tree, so they are not represented as source nodes. Links for unsampled capabilities are limited to documented relationships.
Main entry points: Data.java Procurement data · SAAModel.java SAA / CSAA · Solution.java Solution diagnostics · ContextualWassersteinBoxCcgSolver.java Wasserstein CCG
Components 25
Data and configuration
- History loader
src/Helper/basicHelper/CsvHistoryLoader.java - Sample builder
src/Helper/basicHelper/SampleBuilder.java - Instance generator
src/Helper/basicHelper/InstanceGenerator.java - Run configuration
src/Helper/basicHelper/Config.java - Procurement data
src/Basic/Data.java - Procurement parameters
src/Basic/ProcurementParams.java
Context weighting
- Context weights
src/Helper/calculateHelper/WeightCalculator.java - KNN weights
src/Helper/calculateHelper/KNearestNeighborWeightCalculator.java
Optimization models
- SAA / CSAA
src/Model/SAAModel.java - Recourse evaluation
src/Model/SecondStageEvaluator.java - RCSAA decomposition
src/Model/RCSAALBBDSearchSolver.java - RCSAA support
src/Model/RCSAADecompositionSupport.java - Divergence DRO
src/Model/DROBenders.java - Wasserstein DRO
src/Model/ContextualWassersteinBoxSolver.java - Wasserstein CCG
src/Model/ContextualWassersteinBoxCcgSolver.java - Wasserstein input
src/Model/WassersteinBoxInput.java - Wasserstein oracle
src/Model/WassersteinBoxOracle.java
Python experiments
- Forest weights
analysis/trb_svu/rf_leaf_weights.py - PCM solver
analysis/trb_svu/solve_pcm.py - Feasible solver
analysis/trb_svu/msk_feasible_solver.py
Results
- Solution diagnostics
src/Model/Solution.java - Result writer
src/Helper/basicHelper/OutputManager.java
Other components
- Researcher
- CPLEX
- MOSEK
Connections 29
- Researcher → History loader: loads history
README.md - Researcher → Instance generator: runs experiments
README.md - History loader → Sample builder: builds samples
src/Helper/basicHelper/SampleBuilder.java - Instance generator → Procurement data: creates instances
src/Helper/basicHelper/InstanceGenerator.java - Sample builder → Procurement data: assembles data
src/Helper/basicHelper/SampleBuilder.java - Run configuration → Context weights: configures weighting
src/Helper/calculateHelper/WeightCalculator.java - Procurement data → Context weights: weights samples
src/Helper/calculateHelper/WeightCalculator.java - Procurement data → KNN weights: weights neighbors
src/Helper/calculateHelper/KNearestNeighborWeightCalculator.java - Procurement data → SAA / CSAA: solves scenarios
src/Model/SAAModel.java - SAA / CSAA → Recourse evaluation: evaluates recourse
src/Model/SAAModel.java - Procurement data → RCSAA decomposition: solves robust model
src/Model/RCSAALBBDSearchSolver.java - RCSAA decomposition → RCSAA support: solves scenario cuts
src/Model/RCSAALBBDSearchSolver.java - Procurement data → Divergence DRO: solves DRO
src/Model/DROBenders.java - Procurement data → Wasserstein DRO: solves contextual DRO
src/Model/ContextualWassersteinBoxSolver.java - Procurement data → Wasserstein input: builds input
src/Model/WassersteinBoxInput.java - Wasserstein DRO → Wasserstein input: uses input
- Wasserstein CCG → Wasserstein input: uses input
src/Model/ContextualWassersteinBoxCcgSolver.java - Wasserstein CCG → Wasserstein oracle: separates demands
src/Model/ContextualWassersteinBoxCcgSolver.java - Researcher → Forest weights: runs Python helper
README.md - Researcher → PCM solver: runs PCM experiment
README.md - PCM solver → Feasible solver: uses feasibility solver
analysis/trb_svu/solve_pcm.py - SAA / CSAA → Solution diagnostics: returns solution
src/Model/SAAModel.java - Divergence DRO → Solution diagnostics: returns solution
src/Model/DROBenders.java - Wasserstein DRO → Solution diagnostics: returns solution
src/Model/ContextualWassersteinBoxSolver.java - Result writer → Solution diagnostics: writes diagnostics
src/Helper/basicHelper/OutputManager.java - Result writer → Procurement data: writes data outputs
src/Helper/basicHelper/OutputManager.java - SAA / CSAA → CPLEX: optimizes model
README.md - Wasserstein CCG → CPLEX: optimizes master
src/Model/ContextualWassersteinBoxCcgSolver.java - PCM solver → MOSEK: solves PCM
README.md