Introduction

H eterogeneous

A gent

R ecursive

D ynamic

P ublic

I nsurance

G ames.

This web documentation details the methods proposed in the paper by Kam and Stauber, “Computing Dynamic Public Insurance Games with Endogenous Agent Distributions”.

This is a downloadable PDF version of the documentation.

• 1. State Space
  • 1.1. Properties
• 2. Actions and Transitions
• 3. Intersections with State-space Partitions
  • 3.1. Polytope intersection problems
• 4. Sampling on State-space Partitions
• 5. State Space Computations
  • 5.1. Storage
  • 5.2. Implementation
• 6. Equilibrium Payoff Correspondence
  • 6.1. Background
• 7. Approximating SSE operators
  • 7.1. Conceptual
  • 7.2. Practical
• 8. Bilinear programs and SSE operator
• 9. Punishment values and BLPs
  • 9.1. Implementing punishment values
• 10. Computing approximate SSE payoffs
  • 10.1. Outer Approximation: Conceptual
  • 10.2. Outer Approximation: Implementation
  • 10.3. Inner Approximation: Conceptual

Software

HARDPIG relies on the following software:

• MATLAB platform

• YALMIP BMIBNB global optimization solver

  • Bilinear Matrix Inequality Branch-and-Bound solver

• GNU GLPK linear programming solver (in ANSI C)

• SNOPT general-purpose local optimization solver (Fortran)

  • MATLAB Executable binaries files for up to 300 variables and 300 constraints available freely from Phillip E. Gill.

Source codes (in MATLAB) and executables (in C/Fortran) are available from the authors via email.

Indices and tables

• Index
• Search Page