Defining Random Search Termination Conditions

Abstract

Test generation can be accomplished at different levels: micro-level, gate-level and functional level. Functional test generation is usually based on simulation during which output values are computed for given input stimuli. The problems arising in this context can be solved by random and deterministic search methods for discrete optimization. The aim of this paper is to formulate the test generation problem as an optimization problem and to define random search termination conditions. In many cases the termination condition determines quality of the solution. It is demonstrated that during functional test generation there may be used various random search termination conditions and experimentally evaluated the quality of the obtained solution. The proposed random search termination conditions may be used in solving of other optimization problems. The presented research results enable reasonably to choose the appropriate termination conditions for proper search scope and precision of the solution. Ill. 1, tabl. 4, ref. 8 (in English, Abstracts in English, Russian and Lithuanian).