A Comprehensive Survey on Fitness Landscape Analysis

Evolutionary algorithms and other metaheuristics in water resources: Current status, research challenges and future directions

“…However, formally distinguishing the statistical behaviour of EAs on a given topology is an uncertain process itself that requires rigorous statistical experiments (see Section 2.3 and Back et al, 1997;Bayer and Finkel, 2004;Goh and Tan, 2009;Goldberg, 2002;Hadka and Reed, 2012;Matott et al, 2012). As such, the most common motivation for assessing the problem characteristics through fitness landscape analysis is to gain a better understanding of optimisation algorithm performance on a given set of problems (Pitzer and Affenzeller, 2012). Jones (1995) and Stadler (2002) provide a formal definition of the fitness landscape, where the landscape is defined as a directed graph, where the nodes correspond to solutions, and the edges (connections) to search operator manoeuvres.…”

“…Based on this definition, the fitness landscape is not only dependent upon the problem to be solved, but also the choice of algorithm through its parameterised search operators. In some cases, the landscape defined by an operator is strongly related (or even equivalent) to the topological distance between solutions in the decision space (Moraglio and Poli, 2004), for which cases it may be sufficient to study the fitness landscape as defined by the topology of the fitness function (Pitzer and Affenzeller, 2012). Despite the intimate link between a problem's fitness function and the induced algorithm behaviour (as indicated by the connecting arrow in Figure 2), this section discusses issues associated with the characterisation of the landscape properties (as they depend on the topology defined by algorithm operators), and Section 2.3 focusses on the temporal characterisation of an algorithm's run-time search behaviour.…”

“…A summary of the major characteristics of fitness landscapes that have been identified, and metrics to quantify them, is provided in Table 3. For more information on these characteristics, the reader is directed to two recent reviews on the topic of fitness landscape analysis, including Pitzer and Affenzeller (2012) and Malan and Engelbrecht (2013). Table 3: Fitness landscape characteristics, how they influence problem difficulty, and an example metric to quantify the characteristic.…”

“…While fitness landscape analysis is usually much more resource intensive than just solving a given problem, the resulting insights can be valuable for increasing problem understanding (Pitzer and Affenzeller, 2012). As such, fitness landscape analysis may not be the best use of resources in applied studies where the sole aim is to solve a single problem instance, typically to assist in decision making.…”

Metaheuristics for agricultural land use optimization. A review

“…The fitness landscape analysis, in the metaheuristics community, can be a promising idea for this purpose (He et al 2007). According to Pitzer and Affenzeller (2012), the fitness landscape analysis is commonly used to better understand the performance and the progression "through the landscape" of a given heuristic on a problem instances (Pitzer and Affenzeller 2012). The use of such a concept allows characterizing local optima, plateaus, basins of attraction, and barriers.…”