A library of examples for GGPLAB

The following is a library of examples for GGPLAB, a Matlab-based toolbox for specifying and solving geometric programs (GPs) and generalized geometric programs (GGPs).

Each example comes with its full Matlab code, and an HTML page that lists the output generated by Matlab when the code is run. You need to download your own copy of GGPLAB distribution in order to run these examples on your system.

Here is a simple code segment to get you introduced to the GGPLAB usage.

gpvar x y z % create three scalar GP variablesm1 = 3.4*x^-0.33/z % form a monomial p1 = z*sqrt(m1)+0.1/m1 % form a posynomial gp1 = max(1,x+y,p1) % form a generalized posynomial % form an array of constraints constrs = [ m1==x, p1<=m1, 1<=y, gp1+p1<=5/y ] % solve generalized GP and assign solution to GP variables [obj_value, solution, status] = gpsolve(x+y+z,constrs) assign(solution)

These examples are public domain. You are free to use them in any way you wish; but when you do, we request that you give appropriate credit to the authors.

Last updated: May 22, 2006

GP examples library:

General math problems
- Frobenious norm diagonal scaling (frob_norm_diag_scaling.m)
- Perron-Frobenious norm minimization (min_pf_norm.m)
Structural optimization
- Cantilever beam design (cantilever_beam.m, plot_cbeam.m)
- Cantilever beam design with no recursion (cantilever_beam_norec.m, plot_cbeam.m)
Geometry optimization
- Floor planning (floor_planning.m)
- Maximum volume box (max_volume_box.m)
Machine learning

Logistic regression (log_regression.m, load_log_reg_data.m)

Analog circuits design
Digital circuits design

Semiconductor devices
- Optimal doping profile (basic_odp.m)
- Optimal doping profile with current gain constraint (beta_min_odp.m)
Communications and information theory
- Simple power control in communications system (power_control.m)
- Discrete memoryless channel capacity (dmc_capacity.m, entropy.m)
Finance
- Simple spot rate curve extraction from financial data (simple_spot_rate_extraction.m, aalpha.m, ar.m, b.m)

Back to the GGPLAB home page.