Quadratic Formulation of Mutual Information for Sensor Placement Optimization using Ising and Quantum Annealing Machines

Yuta Nakano,Shigeyasu Uno
2024-10-15
Abstract:We address a combinatorial optimization problem to determine the placement of a predefined number of sensors from multiple candidate positions, aiming to maximize information acquisition with the minimum number of sensors. Assuming that the data from predefined candidates of sensor placements follow a multivariate normal distribution, we defined mutual information (MI) between the data from selected sensor positions and the data from the others as an objective function, and formulated it in a Quadratic Unconstrainted Binary Optimization (QUBO) problem by using a method we proposed. As an example, we calculated optimal solutions of the objective functions for 3 candidates of sensor placements using a quantum annealing machine, and confirmed that the results obtained were reasonable. The formulation method we proposed can be applied to any number of sensors, and it is expected that the advantage of quantum annealing emerges as the number of sensors increases.
Quantum Physics
What problem does this paper attempt to address?
The paper aims to address the sensor placement optimization problem, specifically targeting the selection of a predetermined number of sensor locations from multiple candidate positions to maximize information acquisition with the minimum number of sensors. The main contribution of the paper is the proposal of a new method that transforms the sensor placement optimization problem based on mutual information maximization into a Quadratic Unconstrained Binary Optimization (QUBO) problem. This method allows the use of a quantum annealer to solve the optimal sensor placement scheme. The paper first assumes that the data from candidate sensor locations follow a multivariate normal distribution and defines mutual information as the objective function. Then, the paper details how to express mutual information in QUBO form and achieves this transformation through the calculation of a mask matrix and determinant. To verify the effectiveness of the proposed method, the paper uses two toy models for experimental validation and demonstrates that the proposed method can successfully find the optimal combination in the case of 3 sensors. However, as the number of sensors increases, the problem becomes more complex, requiring further research on how to transform Higher Order Binary Optimization (HOBO) problems into QUBO problems and exploring how to achieve Pareto optimization with more sensors.