Performance Evaluation of Runner–Root Algorithm on CEC 2013 Benchmark Functions

From past four decades, many heuristic algorithms are proposed by researchers for solving complex engineering problems. The main source of inspiration of these algorithms is the intelligence present in different parts of nature. These algorithms are found to be better than traditional optimization algorithms for combinatorial optimization problems. F. Merrikh-Bayat proposed a new heuristic algorithm named runner–root, which is inspired by the function of runners and roots of strawberry and spider plants. F. Merrikh-Bayat had tested performance of the algorithm on standard CEC 2005 benchmark problems. Objective of this paper is to test performance of runner–root algorithm on benchmark problems reported in Congress on Evolutionary Computation (CEC-2013) Technical Report. Results show that RRA gives optimal values for all the test functions for dimension two. RRA performance is satisfactory on dimension five except composite functions. For dimension 10 and 30, RRA is better than GA and CMA-ES. But, RRA is suffering from the problem of curse of dimensionality. Its performance is degraded for dimensions 50 and 100.