Article Preview
TopIntroduction
A lot of problems are brought about by present-day technology and societal and environmental processes which are highly complex and large in dimension and stochastic by nature. Scientists and engineers are always faced up with the analysis, design, and synthesis of real-time problems. To overcome these types of problems model order reduction is required to reduce its complexity. But model order reduction (MOR) of these large-scale linear dynamic systems (LSLDSs) is a challenging and exciting area for scientists, engineers, and researchers. None model order reduction technique (MORT) is suitable for all systems. So it motivates the scientists, engineers, and researchers to invent new MORTs to overcome the complexity of real-time problems. Many MORTs are available in the frequency domain (Singh, Chatterjee & Vishwakarma, 2015) (Singh et al., 2016) (Parmar, Mukherjee, et al., 2007) (Prajapati & Prasad, 2019a) (Prajapati & Prasad, 2019c) (Vishwakarma, 2011) (Vishwakarma & Prasad, 2008) (Sikander & Prasad, 2015b) (Tiwari & Kaur, 2017) (Shamash, 1975a)(Tiwari & Kaur, 2020) (Mukherjee et al., 2005) (Sikander & Prasad, 2017) (Kumari & Vishwakarma, 2019b) (Kumari & Vishwakarma, 2019a) (Kumari & Vishwakarma, 2021) as well as time-domain (Jiang & Chen, 2012) (Shamash, 1975b) (Prajapati & Prasad, 2019b) (Vishwakarma & Prasad, 2014) (Eid & Lohmann, 2008) (Hund & Saak, 2018). MORTs are available not only for single-input single-output (SISO) LSLDSs but also for multi-input multi-output (MIMO) LSLDSs (Singh, 2014) (Vishwakarma & Prasad, 2009) (Prasad, 2000) (Alsmadi et al., 2014). The MOR principle is also useful in the internet of things (IoT) environment (Biswas et al., 2018), in the improvement of agricultural production (Kitouni et al., 2018), and the field of healthcare (Srinivasa et al., 2018).
In the present paper, the authors proposed a MORT technique by mixing the advantages of two MORTs and compared the performance properties of conventional and evolutionary techniques. The denominator of the proposed reduced system (RS) is obtained by the stability equation technique (SET) whereas the numerator is obtained by conventional and evolutionary techniques. The essential properties of RS obtained by proposed conventional and evolutionary techniques are compared using Matlab/Simulink which justifies the efficacy of the proposed evolutionary technique over the conventional techniques.
The SET is one of the popular MORT in the frequency domain. T. C. Cheng and C. Y. Chang proposed a SET (Chen et al., 1979) in 1979, which plays a vital role among the most popular MORTs available in the literature. In the SET, the RS is obtained directly from the pole-zero pattern of the original LSLDS. The RS obtained by SET (Chen et al., 1979) is stable if the original LSLDS is stable. The advantage of the SET (Chen et al., 1979) is that it retains the first two time moments of the original LSLDS. Hence not only transient responses but also steady-state responses for the impulse, step, and ramp inputs between the original LSLDS and RS are matched. The SET is also applicable to the non-minimum phase (Chen et al., 1980a) and fast oscillatory systems (Therapos, 1983).