KPG Run is a cross-platform desktop application for solving various power system optimization problems. It provides a modern UI for running production-grade solvers and reviewing optimization results. KPG Run also exports result files compatible with KPG View.
KPG Run uses the KPG 193 test system dataset and visualizes key optimization results. KPG 193 is a synthetic Korean transmission grid model constructed from publicly available data and intended as a benchmark for studies of the Korean power system.
KPG Run includes four optimization models commonly used in power system operations and planning.
Economic Dispatch (ED): Simplest model for cost-minimized generation dispatch without network constraints. → ED Formulation Details
Unit Commitment (UC): Day-ahead scheduling with generator commitment decisions and temporal constraints. → UC Formulation Details
DC Optimal Power Flow (DC-OPF): Network-constrained dispatch using linearized power flow equations. → DC-OPF Formulation Details
AC Optimal Power Flow (AC-OPF): Nonlinear network-constrained dispatch using AC power flow, capturing voltage magnitudes and reactive power. → AC-OPF Formulation Details
KPG Run supports custom solver scripts written in Julia.
To create a custom solver, use the existing .jl solver scripts as templates. Implement your own optimization model in Julia using JuMP.jl, and run it through KPG Run.
For details on using Julia and JuMP, refer to the resources listed below.
Julia Documentation: docs.julialang.org
JuMP Documentation: jump.dev
Feature | ED | UC | DC-OPF | AC-OPF |
|---|---|---|---|---|
Problem Type | LP/QP | MILP | LP/QP | NLP |
Network Model | None | None | DC | AC |
Time Periods | Single | Multiple | Single | Single |
Commitment Decisions | None | Binary | None | None |
Accuracy | Low | Medium | Medium | High |
Solving Speed | Fastest | Slow | Fast | Medium |
Intuitive GUI with clean, modern design
Real-time solver output streaming to console
Parameter validation before solving
Status indicators for running optimizations
Result summarization with key metrics
Data Source: Test system folder selection for the run.
Problem Setup: Problem type configuration and optimization horizon settings.
Solver: Solver engine selection with solver options (e.g., time limit and output level).
Ipopt (open source)
Gurobi (commercial)
Knitro (commercial)
Model | Ipopt | Gurobi | Knitro |
|---|---|---|---|
ED | ✓ | ✓ | ✓ |
UC | ✗ | ✓ | ✓ |
DC-OPF | ✓ | ✓ | ✓ |
AC-OPF | ✓ | ✓ | ✓ |
Each optimization model must be solved with a solver engine that supports its problem type.
Solver choice may affect runtime and convergence.
Academic users may obtain a free Gurobi academic license.
Chain solvers (UC → OPF) using stored results (planned)
Batch processing for multiple scenarios (planned)
CSV export for external analysis
GeoJSON generation for KPG View visualization
Operating System: macOS 10.15+, Windows 10+ (Linux support planned)
Julia: 1.11.2+
Solvers: Ipopt, Gurobi, Knitro
Memory: 8GB RAM minimum, 16GB recommended
Storage: 2GB for application and data
Launch KPG Run
Choose "KPG 193 v1.5" as the data source
Select "DC-OPF" for the problem type
Set Day to 1
Click Solve button and monitor the console output
View results in the Results tab