Modified Spiral Dynamics Optimization with Damped Coordinate Oscillations

Summary

The article presents a modified Spiral Dynamics Optimization method for continuous optimization in multiple dimensions. It describes limitations of pairing coordinates to form spirals, especially for odd-dimensional problems and functions whose coordinates are independent. Its alternative projects spiral-like motion onto each coordinate using a damped harmonic oscillator, with randomized phase and coordinate-specific amplitude measured from the current point to the best known solution.

As a better global point is found, the method recalculates amplitudes, and candidate points oscillate toward that point with damping. The article reports comparative test results in a histogram and characterizes the method as simple, fast, and controlled by few external parameters. It also notes substantial variability between results and a tendency to become trapped in local optima. The tests are optimization benchmarks, so the document does not establish trading profitability or suitability for any particular market strategy.

Key ideas

• The modification avoids linking dimensions in pairs, which can bias search on some multidimensional functions.
• Each coordinate follows a damped oscillation around the current best solution rather than forming paired-coordinate spirals.
• Oscillation amplitudes are recalculated when a better global point is found and can differ by coordinate.
• The method uses random phase variation while otherwise relying on the initialized population and its search dynamics.
• Reported advantages include simplicity and speed, while variable outcomes and local-optimum trapping are stated limitations.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.