Memetic Algorithms for the MinLA Problem

Abstract

Abstract. This paper presents a new Memetic Algorithm designed to compute near optimal solutions for the MinLA problem. It incorporates a highly specialized crossover operator, a fast MinLA heuristic used to create the initial population and a local search operator based on a fine tuned Simulated Annealing algorithm. Its performance is investigated through extensive experimentation over well known benchmarks and compared with other state-of-the-art algorithms.

A Refined Evaluation Function for the MinLA Problem

“…The test-suite used in this experiment is composed of the 21 benchmarks 2 proposed in [12] and used later in [10,13,14]. In our preliminary experiments, we observed similar results over the set of 21 benchmark instances.…”

“…To tackle the MinLA problem, a number of heuristic algorithms have been developed. Among these algorithms are a) heuristics especially developed for MinLA, such as the multi-scale algorithm [10] and the algebraic multi-grid scheme [14]; and b) metaheuristics such as Simulated Annealing [12] and Memetic Algorithms [13].…”

“…Furthermore, A simple SD algorithm was used, which is not as effective as Simulated Annealing or Tabu Search, to reduce the strong influence of (sophisticated) local search procedures. In this sense, this MA implementation is much simplified comparing with the MA of [13]. Next all the details of our MA implementation are presented.…”

“…For this purpose we have reused the MA reported in [13] and replace in its code the classic evaluation function by the refined function Φ (call this algorithm MAMP-Φ). The resulting code was compiled in the computational platform were compared over a set of well known benchmarks within a basic SD algorithm.…”

“…The resulting algorithm, called MAMP-Φ, was compared with the two best known heuristics: AMG [14] and MAMP [13]. The results obtained by MAMP-Φ are superior to those presented by the previous proposed evolutionary approach [13], and permit to improve on 8 previous best known solutions.…”