Stabilność i zbieżność modelu gradientowego w interaktywnej optymalizacji wizualnej w trybie bezpośrednim

The article describes the stability and convergence criterion of linear gradient internal model used in on-line interactive optimization. This type of optimization, in fact of increasing computing power of modern computers, allows to solve incompletely determined optimization problems. Stability and convergence of the linear gradient model was discused for three cases: when searching point moving from the forbidden into the allowed area, when searching point moving in the allowed area, and whensearching point moving from the allowed into the forbidden area.