Search

This paper introduces a new optimization technique known as Swine Influenza Model based Optimization (SIMBO). It is mimicked from Susceptible-Infectious-Recovered (SIR) models of swine flu. The development of SIMBO follows through treatment (SIMBO-T), vaccination (SIMBO-V) and quarantine (SIMBO-Q) based on probability. The SIMBO variants can be used to optimize complex multimodal functions with improved convergence and accuracy. Firstly, swine flue test based on the dynamic threshold identifies a confirmed case of swine flue. After a confirmed case of swine flue in the community, the susceptible are advised to go for the swine flue vaccination to acquire immunity. The confirmed case of swine flue is quarantined from the population. The suspected cases are treated with antiviral. The amount of antiviral drugs given to individual is dependent on patients with or without complications as well as current health of individual. In SIMBO-V and SIMBO-Q, state of the individual is updated directly through vaccination/quarantine and indirectly through treatment. The nonlinear momentum factors restrict the individuals' treatment and state inside the defined limits without checking the health every day. SIMBO variants can easily be implemented on parallel computer architecture without having over burden or modifications. The SIMBO-T, SIMBO-V and SIMBO-Q are tested with thirteen standard benchmark functions and results are compared with other optimization techniques. The results validate that, the SIMBO variants perform comparably better. The performance of SIMBO variants are evaluated in terms of quality of optima, number of times heating stopping criteria, convergence, Fitness Evaluations (FEs), t-test, statistical parameters and analysis of variance test (ANOVA). A real time application in video motion estimation is also considered by authors to test the efficiency of the SIMBO variants. The results of motion estimation using proposed variants seems to be faster than the published methods by maintaining similar peak signal to noise ratio.