TY - JOUR
T1 - System of systems analytic hierarchy and stochastic optimisation design
AU - Liu, Nan
AU - Manoochehri, Souran
AU - Yu, Chan
N1 - Publisher Copyright:
© 2014 Inderscience Enterprises Ltd.
PY - 2014/1/1
Y1 - 2014/1/1
N2 - A formulated methodology of developing the system of systems analytical hierarchy and optimisation flow order is presented. The design problem is modelled as a multi-level hierarchical optimisation problem. The multi-levels such as, super system level, system level and subsystem level are defined. Based on the idea of system of systems, each system performance criteria can be decomposed into several subsystems according to their individual functions. At the super system level, performance objective function is considered by minimising the variances between specified expected values of performance functions and their target values and uncertainty effects caused by global variables. The different levels are linked by the global variables. At the same level, all the subsystems are connected by the linking variables. The local variables only give contributions to each according local subsystem. A specified way to decide cascading flow order for the multi-level system optimisation decision-making problem introduced to improve convergence speed, by considering the sensitivity analysis of linking variables for each element. The formulated method is applied to the selected system of systems examples, such as, the sensor network example. The optimised results based on methods such as genetic algorithm (GA) are provided.
AB - A formulated methodology of developing the system of systems analytical hierarchy and optimisation flow order is presented. The design problem is modelled as a multi-level hierarchical optimisation problem. The multi-levels such as, super system level, system level and subsystem level are defined. Based on the idea of system of systems, each system performance criteria can be decomposed into several subsystems according to their individual functions. At the super system level, performance objective function is considered by minimising the variances between specified expected values of performance functions and their target values and uncertainty effects caused by global variables. The different levels are linked by the global variables. At the same level, all the subsystems are connected by the linking variables. The local variables only give contributions to each according local subsystem. A specified way to decide cascading flow order for the multi-level system optimisation decision-making problem introduced to improve convergence speed, by considering the sensitivity analysis of linking variables for each element. The formulated method is applied to the selected system of systems examples, such as, the sensor network example. The optimised results based on methods such as genetic algorithm (GA) are provided.
KW - Complex system
KW - Conceptual-design
KW - Decision-making
KW - Large-scale
KW - Multi-disciplinary optimisation
KW - Probabilistic constrains
KW - Sensitivity analytics
UR - http://www.scopus.com/inward/record.url?scp=84907914140&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84907914140&partnerID=8YFLogxK
U2 - 10.1504/IJSSE.2014.064834
DO - 10.1504/IJSSE.2014.064834
M3 - Article
AN - SCOPUS:84907914140
SN - 1748-0671
VL - 5
SP - 114
EP - 124
JO - International Journal of System of Systems Engineering
JF - International Journal of System of Systems Engineering
IS - 2
ER -