Analyzing the Impact of the Low Level Heuristics of a Hyperheuristic for the Master Bay Planning Problem

A hyperheuristic is a high–level algorithm that can be viewed as a planner on a set of heuristics. ln the computational optimization area, the use of hyperheuristics to provide solutions to real–world problems has incremented. In a hyperheuristic approach, the most suitable heuristic is selected and automatically applied at each step of the solution process for a given problem. It is well–known that the effectiveness of low level heuristics could be significantly different for each instance. Some low level heuristics in a hyperheuristic are fitted better for the particular problem instance than others. This is the reason why it is important to study the impact of low level heuristics on a specific problem. In this chapter, we conducted a detailed analysis of the impact of the available set of low level heuristics on the behavior of a perturbative hyperheuristic algorithm. As a study case, we analyze experimentally a hyperheuristic algorithm for solving the Containership Stowage Problem which considers the ship stability in the ship. In particular, we propose three metrics to assess the performance and effectiveness of the low level heuristics.