Analysis of the interactions between genes by systematic gene disruptions and gene overexpressions is an important topic in molecular biology. This paper analyses the problem of identifying a genetic network from the data obtained by multiple gene disruptions and overexpressions in regard to the number of experiments and the complexity of experiments. An experiment consists of simultaneous gene disruptions and overexpressions and the complexity of an experiment is the number of genes disrupted or overexpressed. We define a genetic network as a boolean network and show a series of algorithms which describe methods for identifying the underlying genetic network by such experiments. Some lower bounds on the number of experiments required for the identification are also proved for some cases. (C) 2002 Elsevier Science B.V. All rights reserved.