Parallel Hybrid Algorithm Optimizes Mobile Robot Path Planning Without Manual Tuning

A parallel hybrid metaheuristic optimization algorithm delivers a 2.97% to 265% performance advantage in path planning while reducing computational running time by 18% to 68% compared to competing algorithms. Reporting in the journal Cluster Computing, researchers Rustu Akay and Mustafa Yusuf Yildirim introduced the parallel hybrid improved sine cosine-tabu search algorithm, abbreviated as PHISCTS. The novel computational framework automates interpolation-based trajectory planning by determining optimal data-point configurations across multi-core systems, avoiding the traditional requirement of manual data-point tuning in mobile robot navigation applications.

The research addresses persistent practical challenges within interpolation-based path planning, an established computational problem in robotics that relies on selecting a specific number of data points for curve or line fitting. In this computational domain, the quantity of data points selected directly affects the resulting solution quality.

To eliminate the need for manual parameter selection and tuning, the authors designed a hybrid optimization scheme that combines the distinct strengths of two computational methods: an improved sine cosine algorithm, designated as ISCA, and tabu search, designated as TS. Within this unified structural architecture, the authors configured the two metaheuristics to operate sequentially.

Following the sequential design, the research team configured the hybrid algorithm to run in parallel across multiple processor cores using different numbers of data points. Operating across these multiple cores, the PHISCTS algorithm conducts parallel searches across a predefined candidate set containing varying data-point quantities. Through this parallel multi-core search mechanism, the system evaluates the candidate point configurations simultaneously and directly outputs the specific number of data points that yields the shortest path length alongside its corresponding completed path.

Before evaluating the algorithm on practical mobile robotic navigation problems, the investigators validated its optimization performance on standard mathematical benchmark functions. The proposed algorithm was first tested against the CEC2015 and CEC2022 benchmark function suites.

In the second evaluation stage, the researchers assessed the proposed method directly on the path planning problem by designing simulated testing environments with different difficulty levels. Across these diverse environments of varying complexity, experimental path planning results demonstrated that the proposed PHISCTS algorithm consistently achieved a 2.97% to 265% performance advantage in path planning over other algorithms. Concurrently, the parallel multi-core execution delivered an 18% to 68% reduction in running time compared to other algorithms across the comparative path planning experiments.

The authors concluded that the substantial performance gains and execution time reductions highlight the role of hybrid and adaptive methods in advancing optimization techniques. By coupling sequential metaheuristics with parallel multi-core candidate evaluation, the study demonstrates an effective framework for mobile robot path planning applications that eliminates manual data-point tuning while delivering significant trajectory improvements and faster computational performance across environments with different difficulty levels.

Subject of Research: Parallel hybrid metaheuristic optimization for mobile robot path planning

Article Title: Parallel hybrid metaheuristics for path planning without manual data-point tuning

Article References: Akay, R., & Yildirim, M. Y. (2026). Parallel hybrid metaheuristics for path planning without manual data-point tuning. Cluster Computing, 29(16), Article 930. https://doi.org/10.1007/s10586-026-06742-1

Image Credits: AI Generated

DOI: 10.1007/s10586-026-06742-1

Keywords: mobile robot path planning, metaheuristics, sine cosine algorithm, tabu search, parallel computing, trajectory optimization, interpolation, computational efficiency, Parallel, hybrid, path, planning