Pareto-Based Multi-Criteria Optimization of Risk Management Decisions in IT Projects

Abstract

In practice, risk management in the implementation of IT projects comes down to selecting a set of preventive measures within a limited budget. The difficulty is that no single measure is unambiguously better than another: strengthening quality control reduces the probability of defects but extends the schedule; infrastructure redundancy improves reliability but increases the total cost of ownership; engaging external expertise reduces technological risk while simultaneously creating vendor dependency. The problem is therefore multi-criteria by its very nature. A common way to circumvent this difficulty is scalarisation, that is, the aggregation of all criteria into a single integral indicator with predefined weighting coefficients. Such an approach is computationally convenient but conceals the structure of the trade-off from the decision maker.