Oops! Something went wrong while submitting the form.
Results
Optimization Solver
Optimization Solver
Benchmarking BQPhy®'s QIEO Algorithm: Up to 3.9x Faster Than Genetic Algorithms
Contact Us
Thank you! Your PDF will get downloaded automatically.
Oops! Something went wrong while submitting the form.
Challenges
High-dimensional engineering optimization in digital missions is computationally expensive for traditional classical, gradient, and meta-heuristics algorithms.
Numerous cost function evaluations are required for complex problems which leads to
Convergence at local minima
Suboptimal designs
Inefficient simulations
Results
Implementing QIEO on existing GPU systems enhances parallelization, increases search space exploration capabilities and achieves greater accuracy
QIEO achieved speedups of up to 3.9x across benchmark functions (2.9x for Ackley, 3.9x for Rosenbrock, and 3.84x for Rastrigin).
It required significantly fewer function evaluations (up to 12x fewer for Ackley) and converged up to four times faster than GA (one-fourth the time for Rosenbrock and Rastrigin). Accurate identification of global minima (even if multiple) leading to More optimal design
Akley, Rastrigin, and Rosenbrock Test functions, known as artificial landscapes, were used to evaluate characteristics of BQPhy QIEO such as convergence rate, precision, robustness and general performance