Control algorithms risks in IT projects

In this work basic approaches to project management of IT of a profile in a foreground mode are considered, especially when entering adjustments into structure of projects which undoubtedly has the specific features. The main difference from a stage of forming of the program of project management is that the exception of it can lead to the additional expenses connected with the termination of already begun works, therefore, risks of failure to follow the specified projects increase at the scheduled time. Compensations to performers, penalties for cancellation of contracts and other similar losses belong to such expenses, for example. Task analysis of operational management of the program and a portfolio of IT of projects taking into account these additional expenses is presented in many works. However, for a solution of similar tasks application of any only method is not always possible that sets a task of development of an integrated approach to risk assessment of IT of projects for the choice of the least risky that demands researches in the field of mechanisms of complex estimation. The research objective consists in forming of expert system of support of decision-making at risk assessment of IT of projects taking into account contradictions of knowledge of experts and forming of a compromise in problems of complex estimation and paired comparisons. Research methods. Are applied by projects taking into account risks to a solution of a problem of operational planning of IT: the method of network programming and a method based on the table of admissible solutions. At risk assessment the qualimetrical scale creating three groups forms: low risks, average risks and high risks. At the same time the system of points is appropriated to each group of risks that allows experts to create rather sure estimates. Results. The offered effective algorithm of receiving expert estimates allowed to simplify the procedure of creation of tables of admissible solutions. If the table size (number of the considered solutions) exceeds the set restriction, “compression” of the table to the required size by “pasting” of solutions is made. Pasting of solutions happens in such a way that any admissible solution is not lost (however, inadmissible solutions can appear). As a result we receive upper assessment for an initial problem of risk assessment of IT of projects. Thus, thanks to it is applicable methods of network programming it is possible to exclude the probability of receiving incorrect estimates and uncoordinated opinions of experts at selection of IT of projects in total portfolio of the company of the developer. Conclusion. To receive from experts exhaustive and not contradictory results which will be based on simple and clear actions it is necessary to apply the offered approach to a solution of complex estimation of risks of IT of projects. The first way of accounting of risks (restriction for financing of projects with high and average risks) is effective if value grows with growth of n is polynomial. The second and third ways generally lead to the exponential growth of complexity. However, if the number of projects with high and average risks is limited on top to some number K (which is not depending from n), then both of these ways are effective.