The Random Neural Network with a Genetic algorithm in Intelligent Buildings

Abstract

The Random Neural Network with a Genetic algorithm and its integration into an Intelligent Building: iBuilding is proposed in this paper. The presented biological method, founded on the Genome, codifies and transmits the information from the Intelligent Building. Furthermore, it also multiplexes its data entirely to generate Clusters of Buildings that are interconnected with each other. The key concept proposed in this paper is that the learned information obtained by iBuilding after its interaction with the environment is never lost when the building is decommissioned or retrofitted but transmitted to future iBuilding generations as distributed organisms. Data is codified in the network weights instead of the neurons, similar as the Genome, in order to enable an Artificial Intelligence evolution in iBuilding. The presented biological algorithm is inserted into an iBuilding model where sensorial neurons distributed within the Intelligent Building collect measurements about its environment and select relevant information. This proposed model has been validated with several research datasets that cover several key scenarios; experimental results demonstrate that the Random Neural Network Genetic Algorithm codifies, transmits and multiplexes iBuilding information to future generations with insignificant error, therefore, successfully creating a cluster of buildings.