Lamarckian memetic algorithms: local optimum and connectivity structure analysis

Abstract

Memetic Algorithms (MAs) represent an emerging field that has attracted increasing research interest in recent times. Despite the popularity of the field, we remain to know rather little of the search mechanisms of MAs. Given the limited progress made on revealing the intrinsic properties of some commonly used complex benchmark problems and working mechanisms of Lamarckian memetic algorithms in general non-linear programming, we introduce in this work for the first time the concepts of local optimum structure and generalize the notion of neighborhood to connectivity structure for analysis of MAs. Based on the two proposed concepts, we analyze the solution quality and computational efficiency of the core search operators in Lamarckian memetic algorithms. Subsequently, the structure of local optimums of a few representative and complex benchmark problems is studied to reveal the effects of individual learning on fitness landscape and to gain clues into the success or failure of MAs. The connectivity structure of local optimum for different memes or individual learning procedures in Lamarckian MAs on the benchmark problems is also investigated to understand the effects of choice of memes in MA design. = Conditional probability density function of having offspring y given parent x at generation

Iterated local search with Powell’s method: a memetic algorithm for continuous global optimization

“…A new immigrant scheme is proposed and experimental results on continuous dynamic problems are presented. Le et al [34] introduce the concept of local optimum structure for the analysis of Lamarckian memetic algorithms. They generalize the notion of neighborhood to connectivity structure.…”

Meta-Lamarckian learning in three stage optimal memetic exploration

“…In [4] the classification of coordination methods has been extended and updated. The following four categories have been identified: 1) Adaptive Hyper-heuristic, where heuristic rules are employed (e.g., [8], [9], [10], [11]); 2) MetaLamarckian learning defined in [12], where the activation of the memes depends on their success, see also [13], [14], [15]; 3) Self-Adaptive and Co-Evolutionary, where the rules coordinating the memes are evolving in parallel with the candidate solutions of the optimization problem or encoded within the solution, see [16], [17], [18], [19]; and 4) Fitness DiversityAdaptive, where the activation of the memes depends on a measure of the diversity (e.g., [20], [21], [6], [22], [23]). As a general idea, the algorithmic designer attempts to have a system which performs the coordination automatically.…”

Genetic algorithm with automatic termination and search space rotation

“…Afterward, the exploitation process is invoked through a local search method in order to refine the best candidates obtained so far. GATR thus behaves like a "Memetic Algorithm" [26,34] in order to achieve faster convergence [18,23,37,38].…”