Ising Machines Solve Higher-Order Binary Optimization with Inequality Constraints for Wireless Communications Systems

Wireless communication systems increasingly demand efficient resource allocation, yet optimising these complex networks presents significant computational challenges, particularly when dealing with high-order polynomial relationships and strict limitations. Gan Zheng and Ioannis Krikidis, both from Toshiba, alongside their colleagues, address this problem by developing a novel algorithmic solution using Ising machines to tackle large-scale higher-order binary optimisation. Their work overcomes the limitations of existing quadratic unconstrained binary optimisation approaches, which struggle with the complexity of real-world wireless scenarios, by employing a technique that approximates complex polynomial terms and allows for the efficient solution of a single optimisation problem at each step. This advancement promises to unlock improved performance in wireless networks, as demonstrated through a case study involving reconfigurable intelligent surfaces and simultaneous wireless information and power transfer.

Energy Harvesting Optimization for Wireless Systems

Scientists are tackling complex optimization problems in modern wireless communication, particularly those involving energy harvesting, by combining classical and potentially quantum computing techniques. These problems, often computationally expensive, arise in areas like resource allocation, beamforming, and power control, becoming even more challenging as wireless networks evolve towards 6G. The research explores a hybrid approach, investigating the potential of Quantum Annealing and Simulated Bifurcation as powerful solvers. A key technique involves compressed quadratization, which simplifies complex optimization by transforming higher-order problems into a quadratic unconstrained binary optimization (QUBO) form, suitable for implementation on Quantum Annealers. Researchers also utilize penalty methods to effectively handle inequality constraints within the optimization process, and benchmark their results against established classical algorithms like simulated annealing and the Gurobi optimizer. This work focuses on applications such as Reconfigurable Intelligent Surface (RIS)-aided communication, systems with low-resolution phase shifters, and Wireless Information and Power Transfer (WIPT), aiming to develop methods for formulating complex problems into QUBO form and evaluating the performance of different algorithms on relevant benchmarks.

Taylor Expansion Simplifies Complex Optimization Problems

Scientists have developed a novel algorithm to efficiently solve large-scale higher-order binary optimization (HOBO) problems, crucial for resource optimization in wireless communication systems. Recognizing limitations in applying existing Ising machines to complex wireless scenarios, the team focused on overcoming challenges posed by high-order polynomial terms and strict inequality constraints. Their innovative approach centers on an iterative method based on the augmented Lagrangian method, allowing for the effective handling of constraints during optimization. A key innovation lies in the use of Taylor expansion to approximate complex higher-order polynomials as quadratic ones, simplifying the problem to a single quadratic unconstrained binary optimization (QUBO) problem at each iteration.

This technique streamlines the computational process and enhances efficiency, eliminating the need for auxiliary variables. Researchers demonstrated the method’s efficacy through a case study optimizing phase in a simultaneous wireless information and power transfer system utilizing a reconfigurable intelligent surface with 1-bit phase resolution. Experiments verified that the proposed algorithm achieves satisfactory performance, consistently outperforming benchmark heuristic schemes, demonstrating a significant step towards leveraging the capabilities of Ising machines for complex optimization problems in future wireless communication networks.

Algorithm Optimizes Wireless Power Transfer Systems

This research presents a novel algorithmic solution to address large-scale higher-order binary optimization problems with inequality constraints, a challenge frequently encountered in wireless communication systems. The team successfully adapted the augmented Lagrangian method, combined with Taylor expansion, to approximate complex polynomial terms as quadratic ones, enabling the efficient use of Ising machines for solving these problems. This approach bypasses limitations of existing methods which struggle with the complexity of real-world resource optimization scenarios. The algorithm’s performance was demonstrated through a case study optimizing phase in a simultaneous wireless information and power transfer system utilizing a reconfigurable intelligent surface, confirming that the algorithm achieves satisfactory performance, exceeding the capabilities of benchmark heuristic schemes. This advancement allows for more effective resource allocation and improved system performance in wireless communication networks. While the current implementation focuses on a specific system setup, future research will assess its scalability and adaptability to more complex scenarios, potentially exploring different approximation techniques to further enhance performance and broaden its applicability.