The grid of values can be built around any parameter by explicitly giving the center point of the grid (in response to the model descriptor prompt). If the model descriptor provides only fixed values (as in the example below), the search for the true minimum will start from the grid , and the parameters can vary within `best-fit-parameter ', where is the user input to define the range of the grid (half range of first axis parameter).
input file: rp90_obs.tab
output file: rp90_b_grd.tab
xs> search_grid Performing a grid search. observed spectrum [root_obs.tab]: rp90 model descriptor: ray(0.3 c 100)+ray(0.67 1.26 c 100) root name for output file [root_grd.tab]: rp90_b The chisquare grid will be built around the following model parameters:
These are listed on the screen.
Please choose the grid axes:
model componet number for the first axis (1): <cr> first axis parameter (norm|index|temp|galacticNh|intrinsicNh): temp half range of first axis parameter (0.4): 0.25 number of steps in the first axis (1:) (7): 5 model component number for the second axis (1): 2 second axis parameter (norm|index|temp|galacticNh|intrinsicNh): temp half range of second axis parameter (0.2): 0.5 number of steps in the second axis (1:) (7): 5 Best grid fit found at grid point: 3 3 First parameter value = 3.000E-1 Second parameter value = 1.260E0 Chi-square = 48.181
Final fit model, best fit parameters, will be listed on screen.