No abstract

Graph Coloring, one of the most challenging combinatorial problems, finds applicability in many real-world tasks. In this work we have developed a new artificial bee colony algorithm (called O-BEE-COL) for solving this problem. The special features of the proposed algorithm are (i) a SmartSwap mutation operator, (ii) an optimized GPX operator, and (iii) a temperature mechanism. Various studies are presented to show the impact factor of the three operators, their efficiency, the robustness of O-BEE-COL, and finally the competitiveness of O-BEE-COL with respect to the state-of-the-art. Inspecting all experimental results we can claim that: (a) disabling one of these operators O-BEE-COL worsens the performances in term of the Success Rate (SR), and/or best coloring found; (b) O-BEE-COL obtains comparable, and competitive results with respect to stateof-the-art algorithms for the Graph Coloring Problem.