Constrained optimization (CONOP)

Requires a table of depths/levels, with wells/sections in rows and event pairs in columns: FADs in odd columns and LADs in even columns. Missing events coded with zeros.

PAST includes a simple version of Constrained Optimization (Kemple et al. 1989). Both FAD and LAD of each taxon must be specified in alternate columns. Using so-called Simulated Annealing, the program searches for a global (composite) sequence of events that implies a minimal total amount of range extension (penalty) in the individual wells/sections. The parameters for the optimization procedure include an initial annealing temperature, the number of cooling steps, the cooling ratio (percentage lower than 100), and the number of trials per step. For explanation and recommendations, see Kemple et al. (1989).

Output windows include the optimization history with the temperature and penalty as function of cooling step, the global composite solution and the implied ranges in each individual section.

The implementation of CONOP in PAST is based on a FORTRAN optimization core provided by Sadler and Kemple.

Reference

Kemple, W.G., P.M. Sadler & D.J. Strauss. 1989. A prototype constrained optimization solution to the time correlation problem. In Agterberg, F.P. & G.F. Bonham-Carter (eds), Statistical Applications in the Earth Sciences. Geological Survey of Canada Paper 89-9:417-425.