Fuzzy Forecast Approach to Estimation of Complex Engineering System’s Acceptability

Author Name(s): Ilyas I. Ismagilov, Linar A. Molotov, 3Pavel Zinovev
Author Email: molotov.linar@mail.ru

Abstract

In this paper one of the most actual problem of complex engineering system design and development is discussed. The problem is how to explore and to estimate complex systems possibilities to be reliable and survivable during all its future life cycle. Fuzzy forecasting approach based on group-expertise procedure seems to be most convenient solution to overcome this problem. The new technique to realize worst-case analysis of complex system’s evolution behavior via fuzzy shaping of its acceptability domain is considered. For detailed investigation of acceptability domain evolution way we use composite aggregative criteria for quantitative as well as for qualitative estimation of complex system’s key performance characteristics. As a result of fuzzy forecasting and domain borderline shaping via group-expertise we hope to be entirely convinced that within the system acceptability domain all design specifications, all constructive restrictions and requirements either quantitative or qualitative are successfully satisfied. In conclusion perspective directions for system survivability and acceptability worst-case analysis based on fuzzy forecasting group-expertise procedure are observed.

Introduction

One of the most important and complicated problem has to be solved by contemporary engineering is to design, to construct and to develop various type of complex engineering systems (CES) as reliable and survivable as it possible [15]. During system’s exploitation process CES’s evolution may be either progressive (evolvable) or negative and regressive (degradable). In this way numerous CES’s design problems connected with well-founded worst-case analysis of the complex system evolution and future development trends are very actual and extremely interesting. Nowadays, a system’s dependability assurance during an overall exploitation period may be considered as one of the important constituent of CES’s worst-case design and development process. Functional survivability assurance of CES (performability) is an essential, indefeasible and troublesome part of the general problem [6-9]. Originally, CES’s performability is determined by threshold values of performance quality measures destined to evaluate how successfully our system will carry out all prescribed functions during its life cycle. And CES’s successive realization of all initially specified design functions indicates that all primary system mission goals will be eventually achieved. Many significant CES’s metrics have quantitative nature. But some notable performance metrics (NPM) are qualitative. Hence, there is an urgent necessity to develop some innovative approach to NPM forecasting and appraisement based on multifunctional complex aggregative criteria for quantitative as well as for qualitative estimation and investigation of system’s substantial characteristics. Practical implementation of this comprehensive approach will allow CES’s designers effectively and adequately analyze situations with various system conditions during whole system’s life cycle (SLC). To realize this idea in CES’s design and development process an approach is proposed hereafter which may be employed for the system worst-case acceptability evaluation.

Conclusion

It should be noted that if the set of degradable alternatives will be empty its necessary to make final decision on the base of SAD visual analysis with regard of expert group’s chief individual opinion about worst alternatives which belong to “stable” development trend. Main issues for further research and development will be:  Strengthening of expert group prognoses validity;  Fuzzy neural networks application to support described technique practical realization;  Improving of existing fuzzy shape of SAD after final evaluation of acceptability reserves.

745 total views, no views today

Download PDF File

About the author: admin