Modeling and Analysis of Safety and Risk in Complex Systems

Abstract

The present thematic issue aims to draw attention of the experts in system analysis and design to numerical evaluation of risks and provision of safety in complex systems. Despite the long history of the problem of safety, the number and sizes of accidents (catastrophes) in the world is steadily growing. Many factors are responsible for this, but we are convinced that the main cause lies in insufficient theoretical study of the problem. The Editorial Board of Automation and Remote Control hopes that this issue would stimulate quest for effective solutions of the problems of safety and risk. The methodological aspects of many problems of system safety are allied to those of the reliability theory. However, there are significant differences lying primarily in that the system transitions to dangerous states that are accompanied by serious damages and even human losses are caused by the involved relations of events such as failures of structural system elements, erroneous actions of decision makers, unfavorable natural environmental impact, and so on. Their collections that are often called the accident initiators are defined by the field of application and the operational conditions of a system. That is why, analysis of accident initiation is actually skipped in this issue despite its importance. The papers of the first part of the present issue deal mostly with the subsequent stages of system analysis such as the choice of safety (risk) indices and their numerical evaluation. The methods of their solution are rather universal and can be used to analyze other complex—economic, in particular,— systems. Nevertheless, the methods used for risk analysis and evaluation in economic systems have certain specificity. Beginning from Markovitz, the pioneering works measured the risks of economic systems in terms of the variance (var) of possible incomes, but this risk index proved to be inconvenient for practical purposes. Therefore, the recent economic papers make use of another, more convenient, Value-at-Risk (VaR) index. Mathematically speaking, VaR is the quantile of a corresponding level for distribution of the possible losses. The papers of the second part of this issue are devoted mostly to using this risk index in financial systems. Therefore, with regard for their orientation, the papers of this issue are distributed between the following three parts: