Pareto memetic algorithm for multiple objective optimization with an industrial application

Abstract

Multiple objective combinatorial optimization problems are difficult to solve and often, exact algorithms are unable to produce optimal solutions. The development of multiple objective heuristics was inspired by the need to quickly produce acceptable solutions. In this paper, we present a new multiple objective Pareto memetic algorithm called PMS MO . The PMS MO algorithm incorporates an enhanced fine-grained fitness assignment, a double level archiving process and a local search procedure to improve performance. The performance of PMS MO is benchmarked against state-of-the-art algorithms using 0-1 multi-dimensional multiple objective knapsack problem from the literature and an industrial scheduling problem from the aluminum industry.

Design of an Efficient Genetic Algorithm to Solve the Industrial Car Sequencing Problem

“…If GAs are known to be well suited for multi-objective optimization (Barichard, 2003;Basseur, 2004;Zinflou et al, 2006), few researchers and industrials decided to use this category of algorithms to solve the ICSP. Among the 18 teams that qualified for the second phase of the Challenge, only one proposed a genetic algorithm based approach.…”

Solving the industrial car sequencing problem in a Pareto sense

“…In Zinflou et al [15], we show that the PMS MO algorithm obtains better results than NSGAII [7] and SPEA2 [9] on the multiple objective 0/1 knapsack problem. In the same way, for a real industrial scheduling problem dealing with aluminum casting, the PMS MO [15] algorithm outperforms the NSGAII algorithm [7] according to the different metrics introduced by Zitzler [16] In this paper, we compare the performance of PMS MO [15] and NSGAII [7] while solving the industrial car-sequencing problem introduced by Renault during the 2005 ROADEF Challenge [17] [18].…”

“…Among non elitist approaches which include no specific mechanisms allowing the memorization of the best solutions found during the execution of the algorithm, we may cite MOGA [4], NPGA [5] and NSGA [6]. Among elitist approaches which include one or several mechanisms allowing the memorization of the best solutions found during the execution of the algorithm, one may cite NSGAII [7], SPEA [8], SPEA2 [9], PAES [10], M-PAES [11], PESA [12], PESAII [13], Micro-GA [14] and PMS MO [15].…”

Generation of Business Process Reference Model Considering Multiple Objectives

“…Multi-objective genetic algorithms (MOGA) have been widely used in many areas including order allocation (Wu et al, 2012), scheduling problem (Zinflou et al, 2008;Minella et al, 2008), human resource planning (Abboud et al, 1998;Hajri-Gabouj, 2003) or design and planning problem (Zhou et al, 2003;Gen et al, 2009). Deb et al (2002) used genetic algorithm to solve multi-objective problems and develop non-dominated sorting and further brought up the idea of using the elitist strategy may have a better convergence when using the genetic algorithm (GA).…”