Sensors Placement Optimization in Smart Broiler Chicken Farm for Indoor Environment Monitoring

In smart broiler chicken farms, the internal environment of the farm depends on the weather seasons and position of various actuators, such as ceiling fans, wet curtains, and heaters. Consequently, numerous sensors are indispensable to effectively monitor the diverse aspects of the farm environment. The selection of optimum locations for sensors from a large pool of sensors involves the spatial estimation representation that closely approximates the performance of all the sensors while minimizing the number of sensors. In the proposed work, 25 Internet of Things data sensing nodes are installed at random locations with the experience of farmers. With the objective of sensors’ location optimization, a novel approach is proposed, integrating Bayesian optimization (BO) in conjunction with the novel ordered common elements crossover genetic optimization algorithm, referred to as Bayesian genetic algorithm (BGA). The primary objective of the BGA is to maximize the fitness function, which measures the gray correlation coefficient (GCC) between the reference trend and the optimal sensor combination. BO is employed to determine the hyperparameters of the genetic optimization algorithm within the specified search space. The proposed solution introduces a novel crossover method for GA known as ordered common elements crossover. The proposed BGA applied to seasonal data to consider the seasonal effects in the sensor location optimization. The effectiveness of this proposed BGA method is compared with the conventional one-point crossover-based BGA. The proposed method has a low value of mean absolute error (MAE) compared to the existing one-point crossover-based BGA, indicating its superior performance.