Chaotic Optimization with Adaptive Search and Weighted Gradients

Summary

This article outlines an enhanced Chaos Optimization Algorithm intended to search complex, multimodal objective functions. Its three phases combine a broad search driven by chaotic maps, a weighted-gradient search that accounts for constraints, and a local search around promising solutions. The implementation adds logistic, sinusoidal, and tent maps, Latin hypercube initialization, agent velocities, mutation, stagnation handling, adaptive penalties, and a narrowing search radius with occasional long jumps.

The document explains the algorithm’s design and implementation in MQL5, including agent state and methods for updating positions and evaluating candidates. It argues that chaotic sequences can help explore the search space and reduce stagnation, but provides no test results or measured comparison in this installment; evaluation is deferred to a later article. The approach is a general optimization method, so its usefulness for trading depends on the objective, constraints, and validation process to which it is applied.

Key ideas

• The method alternates global chaotic exploration, weighted-gradient movement, and local refinement.
• Multiple chaotic maps and diverse initialization schemes are used to increase search coverage.
• Adaptive penalties and search parameters respond to constraint violations and search progress.
• Velocity, mutation, stagnation resets, and occasional long jumps support agent movement.
• The article describes implementation but leaves performance testing and comparative results to a later installment.

Tags

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