Hawks are solitary animals that spend most of their life alone, with the exception of mating season, and don’t hunt in groups or pairs, with the exception of a peculiar species – the Harris hawks, birds of prey that live predominantly in the Americas and are known for hunting in groups and being able to capture larger prey due to their cooperation instinct.
Their technique is called ‘surprise pounce’ and consists of surrounding escaping prey from the sky and forcing it to run in a zig-zag movement that inevitably pushes the animal between the claws of one of the hawks to then be shared with all the group’s members.
Inspired by this scheme, an international group of scientists developed, in 2019, the so-called Harris Hawks Optimization (HHO) tool that, despite being relatively new, has already raised enormous interest within the global scientific community and has found application in a wide variety of fields.
The HHO sort of reproduces the hunting behavior of the Harris hawks by mimicking their exploring, exploiting and attacking strategy. It works in several steps, as follows: it initializes the hawk population and calculates fitness value for each hawk before selecting the best one; it identifies the position of a rabbit, or rabbit population, and calculates jump and energy strength for each of the rabbits; it then performs exploration and exploitation phases and calculates the rabbit’s best location and fitness.
According to most of the scientists involved, this nature-inspired algorithm exhibits smooth transitions between exploration and exploitation while providing competitive results to complex problems.
With this in mind, a group of scientists from the Babol Noshirvani University of Technology, in Iran, and the University of Guelph, developed a modified version of the algorithm, which they called Whippy Harris Hawks Optimization (WHHO), with the intent of applying it to estimate the model parameters of a PV installation.
“In the proposed algorithm, modifications are applied to the process of [the] hawks’ movement towards the prey,” the researchers explained. “The applied changes reduce the likelihood of getting stuck in local minima and increase the convergence speed, compared to the original HHO.” They added an elimination period, during which a certain number of the worst solutions are eliminated and replaced by new solutions in the search space.
The tool was used to estimate the model parameters of systems based on both mono and polycrystalline modules, considering the effect of temperature and irradiance changes. The main goal of the algorithm, the research team emphasized, is to minimize the amount of the obtained Root Mean Square Error (RMSE). This is a standard way to measure the error of a model in predicting quantitative data. “It was observed that the saturation current of the diode and the photocurrent slightly change with the variation of the temperature and irradiance, respectively, and the rest of the parameters remain almost constant under various operating conditions,” they also stated.
According to their findings, exposed in the paper Parameter estimation of PV solar cells and modules using Whippy Harris Hawks Optimization Algorithm, published in Energy Reports, the achieved results confirm the high reliability of the proposed algorithm and its improved convergence speed compared to other recent optimization methods.
