Adapting Tabu Search with Sector Memory for Continuous Optimization

Summary

This article proposes a continuous-space variant of Tabu Search for optimization problems. It partitions each parameter range into sectors and tracks successful and unsuccessful outcomes with separate white and black counters. New candidate coordinates are sampled probabilistically, favoring sectors with positive history while using negative history to encourage moves away from unproductive areas. The method also describes population-based agents, fitness comparisons, sharing coordinates from the best-known solution, and updates to sector memory over iterations.

The approach adapts ideas from tabu memory, intensification, and diversification to avoid cycling while exploring a large search space without assigning a label to every possible solution. The article presents algorithm mechanics and test-result figures, but the supplied text gives no specific comparative scores or detailed experimental protocol. Its stated limitation is that convergence accuracy could be better; the proposed method is presented as a promising, tunable optimization approach, not as a demonstrated trading strategy.

Key ideas

• The method divides each continuous parameter range into sectors and tracks outcomes separately for each sector.
• Positive sector history guides sampling toward promising regions, while negative history encourages diversification.
• Agents update sector counters by comparing current fitness with the previous iteration.
• The approach combines tabu-style memory with population-based search and can share coordinates from the best-known solution.
• The article notes that convergence accuracy remains a limitation and does not establish trading performance.

Tags

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