Reliable parameter estimation of nonlinear chaotic systems and PMSMs with the stellar oscillation optimizer

Why this matters for real-world machines

From weather prediction to electric cars, many modern technologies rely on mathematical models of how complex systems behave. These models are only as good as the numbers, or parameters, that define them. In chaotic systems—where tiny changes can cause huge effects—finding the right parameters is notoriously difficult. This paper presents a new way, inspired by the rhythmic pulsing of stars, to lock onto those crucial numbers with remarkable reliability, both in abstract chaos models and in a widely used electric motor.

How star-like rhythms guide the search

The authors build on a recent optimization method called the stellar oscillation optimizer. Imagine many candidate solutions behaving like stars that gently oscillate in a mathematical space, shifting their positions in response to how well they perform. Instead of following a simple, repetitive motion, these “stars” combine guidance from the best performers, interactions between pairs of candidates, and a gradually shrinking oscillation amplitude. Early on, the search jumps widely to explore; later, it settles into finer, more careful adjustments around the most promising region. This structure is designed to avoid getting stuck too soon while still homing in on a single high-quality solution.

Turning model matching into a single score

To test this star-inspired searcher, the authors cast parameter estimation as a straightforward matching game. They start with a reference system—either a famous chaotic model or a motor model—and generate a time series of its behavior. For any guessed set of parameters, they simulate the same system again and measure how far the simulated trajectory deviates from the reference at many time points. All those differences are rolled into a single cost value: the smaller the cost, the better the guess. The optimizer’s job is to adjust the parameters until this mismatch score can no longer be meaningfully reduced, given the limits of computer arithmetic.

Putting chaos and motors to the test

The method was challenged on three classic chaotic systems—Lorenz, Chen, and Rössler—that are famous for their extreme sensitivity to initial conditions, as well as on a simplified but realistic model of a permanent magnet synchronous motor, a workhorse in electric drives and industrial automation. For fairness, the authors used the same cost function, numerical integration scheme, population size, and iteration budget across all tests. They compared the stellar oscillation optimizer with several recent nature-inspired algorithms, including methods based on electrical circuit laws, horse racing strategies, animal behavior, and human hiking analogies. Each algorithm was run many times independently to probe not just peak performance but also how consistently it could deliver.

How well the new method performs

Across all four systems, the stellar oscillation optimizer repeatedly drove the mismatch between simulated and reference behavior down to the limits of double-precision arithmetic—essentially as low as a computer can meaningfully represent. In chaotic cases, it recovered the true parameters with vanishingly small error in every run, where competitors often showed more variation or needed more iterations to converge. On the motor model, it again reached the correct parameters with nearly identical results across trials, while also tending to run faster than the alternative methods. Statistical tests confirmed that these advantages were not flukes: the distribution of outcomes for the new optimizer was consistently and significantly better than those of the other algorithms.

What this means in simple terms

In plain language, the study shows that a search strategy modeled after the oscillations of stars can be an exceptionally steady “tuning dial” for complex dynamical systems. In ideal, noise-free simulations, it finds parameter settings that make the model’s behavior indistinguishable from the original system, and it does so reliably from run to run. The authors stress that such near-perfect numbers are not guaranteed in messy real-world measurements, where noise and modeling errors play a role. Still, the results strongly suggest that this stellar-inspired optimizer is a powerful new tool for building accurate, trustworthy models of chaotic processes and practical machines like electric motors.

Citation: Ekinci, S., Izci, D., Jabari, M. et al. Reliable parameter estimation of nonlinear chaotic systems and PMSMs with the stellar oscillation optimizer. Sci Rep 16, 11564 (2026). https://doi.org/10.1038/s41598-026-41940-2

Keywords: chaotic systems, parameter estimation, metaheuristic optimization, electric motor modeling, stellar oscillation optimizer