Economics 623

Some Code for Solving Various Problems

C/Gauss code for computing an inverse of a Normal CDF (also illustrates foreign language importing, via DLL (dynamic link library) of C code into Gauss, as if the external code was a Gauss procedure (the C code is typically orders of magnitude faster than Gauss) C/Matlab code for doing multidimensional interpolation/extrapolation of functions (also illustrates Matlab's MEX facility for importing foreign language code, i.e. in this case C-code) which similar to Gauss results in huge speedups compared to writing the routines in Matlab, especially when there are many nested do-loops involved. `ergodic.g` Gauss code to compute the invariant distribution of an nxn Markov transition probability matrix `quadpoints.g` Gauss code to compute weights and abscissae for Gaussian quadrature `vint.g` Gauss code to interpolate a multidimensional function via simplicial interpolation (see also `vint.c` the C version of this same code) `evp.c` C code to compute the conditional expectation of a continuous function of 2 variables with respect to one of these variables (i.e. with respect to the stochastic exogenous state variable) conditional on the value of the other endogenous state variable. `evpp.c` C code to to the same conditional expectation as evp.c expect this code also returns probabilities that represent the conditional expectation as an expectation over the value function at the N grid points in two-dimensional space, EV= \sum_{i=1}^N V_i p_i(p,q) where V_i are the values of the value function at the grid points, and p_i(p,q) are the probability weights returned by evp when called at state point (p,q) `kseq.g` Gauss code to compute the index of an element in a multidimensional lattice where the lattice is linearly ordered as a 1-dimensional vector according to a multilevel do-loop that steps through each element in the array `kseq.c` C version of `kseq.g` used in `vint.c` above. `mysort.c` C function to sort a vector of numbers from low to high, returning both the sorted vector and the indices of the sorted array (e.g if the vector is 4.2 -1.2 3.2 then the sorted array is -1.2 3.2 4.2 and the sorted indices are 1 2 0, i.e. 4.2, the 0-th element of the array is )

Send questions/comments to: `jrust@econ.umd.edu`

C/Gauss code for computing an inverse of a Normal CDF (also illustrates foreign language importing, via DLL (dynamic link library) of C code into Gauss, as if the external code was a Gauss procedure (the C code is typically orders of magnitude faster than Gauss)
C/Matlab code for doing multidimensional interpolation/extrapolation of functions (also illustrates Matlab's MEX facility for importing foreign language code, i.e. in this case C-code) which similar to Gauss results in huge speedups compared to writing the routines in Matlab, especially when there are many nested do-loops involved.
ergodic.g Gauss code to compute the invariant distribution of an nxn Markov transition probability matrix
quadpoints.g Gauss code to compute weights and abscissae for Gaussian quadrature
vint.g Gauss code to interpolate a multidimensional function via simplicial interpolation (see also vint.c the C version of this same code)
evp.c C code to compute the conditional expectation of a continuous function of 2 variables with respect to one of these variables (i.e. with respect to the stochastic exogenous state variable) conditional on the value of the other endogenous state variable.
evpp.c C code to to the same conditional expectation as evp.c expect this code also returns probabilities that represent the conditional expectation as an expectation over the value function at the N grid points in two-dimensional space, EV= \sum_{i=1}^N V_i p_i(p,q) where V_i are the values of the value function at the grid points, and p_i(p,q) are the probability weights returned by evp when called at state point (p,q)
kseq.g Gauss code to compute the index of an element in a multidimensional lattice where the lattice is linearly ordered as a 1-dimensional vector according to a multilevel do-loop that steps through each element in the array
kseq.c C version of kseq.g used in vint.c above.
mysort.c C function to sort a vector of numbers from low to high, returning both the sorted vector and the indices of the sorted array (e.g if the vector is 4.2 -1.2 3.2 then the sorted array is -1.2 3.2 4.2 and the sorted indices are 1 2 0, i.e. 4.2, the 0-th element of the array is )

Send questions/comments to: jrust@econ.umd.edu