Quantum Neural Network Application for Weather Forecasting

Abstract

Abstract Weather forecasts are made by collecting quantitative data about the current state of the atmosphere and using scientific understanding of atmospheric processes to project how the atmosphere will evolve. This research examines and analyzes the use of Quantum neural networks as a forecasting tool. Specifically, neural network’s features such as parallel distributed processing, self-learning and fault tolerance are explored and the idea is then to combine Quantum Computation with the Neural Networks producing Quantum Neural Networks. Quantum computing is a proposed means of using quantum-mechanical effects to achieve efficient computation. The potential power of a quantum computing is in the superposition of states, allowing exponentially many computations to be done in parallel. During a calculation, the bits (called qubits) that are being manipulated are never in a definite one or zero state, instead can be thought of as being both a one and a zero simultaneously, which allows quantum neural networks to explore many solutions at the same time. While only briefly discussing neural network theory, this research determines the feasibility and practicality of using Quantum neural networks as a forecasting tool for the weather system. From the simulation results, it can be seen that the proposed model produces a reasonable accuracy in training which is conducted with 5 parameters (100 days) of the recent temperature, dew point, humidity, sea level pressure, wind speed with historical data. Once the model is trained, it is used to forecast the weather data for next time period, each time forecasting one point ahead.