Pareto genetic path planning hybridized with multi-objective Dijkstra's algorithm

This paper presents an evolutionary approach to multi-objective path planning. The paths are defined on continuous scenes with disjoint and/or non-convex obstacles, for robots moving towards their destinations along linearly piecewise trajectories with any number of vertices. The fastest feasible route is genetically selected via a simultaneous minimization of path length and path steering angle. In order to assure an effective partial sorting of the potential solutions, the genetic algorithm makes use of a self-adaptive Pareto ranking scheme, based on individuals' grouping and dominance analysis. Before ranking, all the unfeasible solutions are corrected, by replacing the sub-paths which intersect the obstacles. The feasible segments are chosen from graphs generated with Delaunay triangulation, by applying a new multi-objective ranking-based fastest path procedure derived from the classical Dijsktra's algorithm. This new procedure is compatible with any nonlinear objective functions and allows using the same ranking scheme during the evolutionary search and correction of paths, thus making possible to guarantee the feasibility of the paths without any significant intrusion in the evolutionary exploration. The proposed algorithm can also be used for solving any graph - based path planning problem, involving any number of objectives. The experiments show the usefulness of the suggested techniques on working scenes with different layouts of obstacles.

[1]  Oscar Castillo,et al.  Multiple Objective Genetic Algorithms for Path-planning Optimization in Autonomous Mobile Robots , 2006, Soft Comput..

[2]  S. Travis Waller,et al.  Parametric search and problem decomposition for approximating Pareto-optimal paths , 2012 .

[3]  Gary B. Lamont,et al.  Evolutionary algorithms for solving multi-objective problems, Second Edition , 2007, Genetic and evolutionary computation series.

[4]  Lily Rachmawati,et al.  Multiobjective Evolutionary Algorithm With Controllable Focus on the Knees of the Pareto Front , 2009, IEEE Transactions on Evolutionary Computation.

[5]  Fumitoshi Matsuno,et al.  Path planning for an autonomous mobile robot considering a region with a velocity constraint in a real environment , 2012, Artificial Life and Robotics.

[6]  Ferariu Lavinia,et al.  Multi-objective genetic algorithm with clustering-based ranking and direct control of diversity , 2013, 2013 17th International Conference on System Theory, Control and Computing (ICSTCC).

[7]  G. G. Stokes.  "J." , 1890, The New Yale Book of Quotations.

[8]  Matthias Ehrgott,et al.  A comparison of solution strategies for biobjective shortest path problems , 2009, Comput. Oper. Res..

[9]  Kalyanmoy Deb,et al.  Multi-objective optimal path planning using elitist non-dominated sorting genetic algorithms , 2012, Soft Computing.

[10]  Gary B. Lamont,et al.  Evolutionary Algorithms for Solving Multi-Objective Problems , 2002, Genetic Algorithms and Evolutionary Computation.

[11]  Marina L. Gavrilova,et al.  Density-Based Clustering Based on Topological Properties of the Data Set , 2008, Generalized Voronoi Diagram.

[12]  Mansoor Davoodi Monfared,et al.  Multi-objective path planning in discrete space , 2013, Appl. Soft Comput..

[13]  Lawrence Mandow,et al.  A comparison of multiobjective depth-first algorithms , 2013, J. Intell. Manuf..

[14]  Faruk Polat,et al.  Limited-Damage A*: A path search algorithm that considers damage as a feasibility criterion , 2011, Knowl. Based Syst..

[15]  Corina Cimpanu,et al.  Multiobjective hybrid evolutionary path planning with adaptive pareto ranking of variable-length chromosomes , 2014, 2014 IEEE 12th International Symposium on Applied Machine Intelligence and Informatics (SAMI).

[16]  Gang Chen,et al.  Preserving and Exploiting Genetic Diversity in Evolutionary Programming Algorithms , 2009, IEEE Transactions on Evolutionary Computation.

[17]  Marina L. Gavrilova,et al.  Generalized Voronoi Diagram: A Geometry-Based Approach to Computational Intelligence , 2008, Generalized Voronoi Diagram.

[18]  Joshua D. Knowles,et al.  Modes of Problem Solving with Multiple Objectives: Implications for Interpreting the Pareto Set and for Decision Making , 2008, Multiobjective Problem Solving from Nature.