G. T. Parks and I. Miller. Selective Breeding in a Multiobjective Genetic Algorithm. Lecture Notes in Computer Science, 1498:250--259, 1998.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....other hand, DLTZ6 tests an MOEA s ability to converge to a curve. In this case there is only one independent variable describing the PO front. 3 Algorithms Used Earlier MOEAs (MOGA [10] NSGA [17] and NPGA [12] were critisized for their dependence on sharing parameter [8] and lack of elitism [15, 18]. Different algorithms that over come these shortcomings have been proposed. Few such algorithms are PAES [13] SPEA [21] and NSGA II [8] In these algorithms, elitism maintains the knowledge acquired during the algorithm execution by conserving the individuals with best fitness in the ....

G. T. Parks and I. Miller. Selective Breeding in a Multiobjective Genetic Algorithm. In A. E. Eiben, M. Schoenauer, and H.-P. Schwefel, editors, Parallel Problem Solving From Nature -- PPSN V, pages 250-259, Amsterdam, Holland, 1998. Springer-Verlag.

.... work with a population of points that is crucial to find the non dominated solution set [1] In this paper, the Nondominated Sorting Genetic Algorithm NSGA [2] 3] is examined with respect to different kinds of elitist techniques, that is, standard, clustering [4] and Parks Miller [5]. Elitism is nowadays a recognized approach to improve the performance of evolutionary based algorithms. This type of approach in multiobjective optimization is not so simple as in single objective problems. However, in multiobjective evolutionary algorithms MOEAs, elitist techniques are used to ....

Parks, G. T. and I. Miller, "Selective breeding in a multiobjective genetic algorithm". In A. E. Eiben, T. Bck, M. Schoenauer and H. Schwefel (Eds.), Fifth Int. Conf. on Parallel Problem Solving from Nature, pp. 250-259. Springer.

....use, as fitness sharing as a tool to keep diversity in the population through the whole Pareto front and hence they require a fitness sharing factor to be set. 2. They did not incorporate elitism, which has been demonstrated to improve significantly the performance of multi objective algorithms [PM98, ZDT00] 3.1. ELITIST AND SHARING PARAMETERLESS MULTI OBJECTIVE EVOLUTIONARY ALGORITHMS Elitism maintains the knowledge acquired during the algorithm execution by conserving the individuals with best fitness in the population or in an auxiliary population. Some algorithms that make use of both ....

Geoffrey T. Parks and I. Miller. Selective Breeding in a Multiobjective Genetic Algorithm. In A. E. Eiben, M. Schoenauer, and H.-P. Schwefel, editors, Parallel Problem Solving From Nature --- PPSN V, pages 250--259, Amsterdam, Holland, 1998. Springer-Verlag.

.... handled as objectives and the resulting problem is solved by the Niched Pareto Genetic Algorithm NPGA [1] Handling constraints as objectives was recently presented for single objective optimization [2] The original NPGA was modified by incorporating the Parks Miller elitist technique (P M) [3], which needed some changes when constraints were treated as objectives. The required changes were essential to avoid convergence toward an infeasible space. Two analytical test problems, TBU [4] and CPT7 [5] that were designed with special features to difficult the PO front search, are chosen ....

....the front defined in the region III. On the other hand, when handling constraints as objectives region II is not viewed as a wall because points of region III do not dominate those of II. IV. MODIFIED PARKS MILLER ELITISM To improve the NPGA performance the Parks Miller elitist technique [3] was used. It consists in incorporating the efficient individuals of the on line population (Pon) to the off line population (Poff) at each generation. When Poff size exceeds a threshold, the dominance criterion is applied, eliminating all dominated solutions. If Poff size continues bigger than ....

G. T. Parks and Miller, `Selective breeding in a multiobjective genetic algorithm'. A.E. Eiben, M.Schoenauer, and H.-P. Schwefel, editors, Parallel Problem Solving From Nature PPSN V, pp. 250-259, Holland, Springer-Verlag, 1998.

.... available, given a set of constraints (for instance legal requirements and size limits of the product) In [1] for example, multi objective optimisation is applied to four performance measures of a gas turbine, in [2] different loads in trusses are the competing objectives to be miniraised and in [3] different properties of a pressurised water reactor load pattern are optimised. The curve (for two objectives) or surface (more than two objectives) that describes the optimal trade off possibilities between objectives is known as the Pareto front [4] A feasible solution lying on the Pareto ....

....[20, 21, 22] These problems also occur in SPEA, PAES and other existing MOEAs which use a truncated elite archive. A remedy to this situation is simply to retain all the non dominated solutions found (as an active input to the continuing search process) as, for example, used by Parks and Miller [3]; however, this approach can be very time consuming (as any individual inserted into the elite archive must first be compared to every individual already present in the archive) It is important to note that in a large number of studies an elite offiine store of solutions is maintained which is ....

[Article contains additional citation context not shown here]

G. T. Parks and I. Miller. Selective Breeding in a Multiobjective Genetic Algorithm. Lecture Notes in Computer Science, 1498:250-259, 1998.

....the true Pareto front. The goal, therefore, of multi objective algorithms (MOAs) is to locate the Pareto front of these non dominated solutions. Multi Objective Evolutionary Algorithms (MOEAs) are a popular approach to confronting these types of problem by using evolutionary search techniques [1, 4, 7, 5, 9, 8, 10, 12, 13, 17, 16, 19, 30, 20, 22, 24, 26, 27, 29, 31, 28]. The use of Evolutionary Algorithms (EAs) as a tool of preference is due to such problems being typically complex, with both a large number of parameters to be adjusted, and several objectives to be optimised. EAs, which can maintain a population of solutions, are in addition able to explore ....

G. T. Parks and I. Miller. Selective Breeding in a Multiobjective Genetic Algorithm. Lecture Notes in Computer Science, 1498:250-259, 1998.

....where a single individual with the current best objective function value is inserted into the next generation. In multiobjective applications some fraction of the solutions along the current nondominated front must be passed on to the next generation (see Reed et al. 2001, Zitzler Thiele 1999, Parks Miller 1998, Bck 1996, and Ishibuchi Murata 1996 for a description of alternative multiobjective elitist strategies) Elitism has been shown to improve the performance and convergence of the GA in both single and multiple objective applications (Thierens et al. 1998, Zitzler et al. 2000, Reed et al. ....

Parks, G. T. & Miller, I. (1998) Selective breeding in a multiobjective genetic algorithm. In A. E. Eiben, T. Bck, M. Schoenauer, and H. Schwefel, eds. Fifth International Conference on Parallel Problem Solving form Nature, pp. 250-259, Springer: Berlin, Germany.

.... an EMOO technique (see for example [14, 6] The main emphasis has been on using an external le that stores nondominated vectors found during the evolutionary process which are reinserted later in the population (this can be seen as a form of elitism in the context of multiobjective optimization [10, 17, 22]) Following the same line of thought of this current research, we decided to develop an approach in which we would use a GA with a very small population size and a reinitialization process (the so called micro GA) combined with an external le to store nondominated vectors previously found. ....

Georey T. Parks and I. Miller. Selective Breeding in a Multiobjective Genetic Algorithm. In A. E. Eiben, M. Schoenauer, and H.-P. Schwefel, editors, Parallel Problem Solving From Nature | PPSN V, pages 250-259, Amsterdam, Holland, 1998. Springer-Verlag.

....the capability of EMO algorithms to approximate the set of optimal trade offs in a single optimization run. These approaches did not incorporate elitism explicitly, but a few years later the importance of this concept in multiobjective search was recognized and supported experimentally 1 (Parks and Miller 1998; Zitzler, Deb, and Thiele 2000) A couple of elitist multiobjective evolutionary algorithms were presented at this time, e.g. SPEA (Zitzler and Thiele 1998; Zitzler and Thiele 1999) and PAES (Knowles and Corne 1999) SPEA, an acronym for Strength Pareto Evolutionary Algorithm, was among the ....

Parks, G. T. and I. Miller (1998). Selective breeding in a multiobjective genetic algorithm. In A. E. E. et al. (Ed.), Parallel Problem Solving from Nature -- PPSN V, Berlin, pp. 250--259. Springer.

....more that just one aspect, it is very difficult to identify the features which are mainly responsible for the better performance of one algorithm over another. On the contrary, a few other studies take one algorithm and focus on a specific operator or parameter to tune, e.g. the selection method [10, 12]. In this case the results are valid for the algorithm under concern and highly dependent on the other algorithmic parameters. Hence, it has remainedopenuptonow how a certain parameter or a certain operator affects the overall performance independent of the specific implementation and the ....

....during the run, but in most cases it just contains non dominated solutions and therefore approximates the Pareto set. If the archived solutions reproduce as well, we say the algorithm uses elitism. Recent studies suggest that the use of elitism improves multi objective evolutionary algorithms [10, 12, 18]. Another issue is the assignment of fitness values to the individuals. In this study we concentrate on Pareto based methods because of their acknowledged advantages over aggregation and population based methods [5] Different techniques inferring a scalar value from the partially ordered ....

[Article contains additional citation context not shown here]

G. T. Parks and I. Miller. Selective breeding in a multiobjective genetic algorithm. In A. E. Eiben et al., editor, Parallel Problem Solving from Nature -- PPSN V, pages 250--259, Berlin, 1998. Springer.

.... this view is supported by sufficient comparison with other techniques or not, there have certainly been a large number of successful applications of MOEAs to real world problems (see [Coe99b] for a comprehensive list) and a steady stream of variations on the theme [HL92, FF93, HNG94, SD94, BW97, PM98] since the pioneering work of Schaffer [Sch84] Recently, a large subset of the well known MOEA schemes was compared [ZDT99] using a set of test functions of differing difficulty and encompassing problem features hypothesized to cause difficulties to multiobjective optimizers, provided by Deb ....

Geoffrey T. Parks and I. Miller. Selective Breeding in a Multiobjective Genetic Algorithm. In A. E. Eiben, M. Schoenauer, and H.-P. Schwefel, editors, Parallel Problem Solving From Nature --- PPSN V, pages 250--259, Amsterdam, Holland, 1998. Springer-Verlag.

.... Strategy Knowles and Corne, 2000) Both PAES and SPEA have been shown to outperform sophisticated versions of NPGA and NDS on a variety of benchmark problems, while various other modern MOEAs exist which have been shown to perform well on particular applications (eg: Fonseca and Fleming, 1995; Parks and Miller, 1998), but have not yet been systematically compared against other modern MOEAs on a common set of test problems. Somewhat removed from the MOEA research community, researchers in multiple criteria decision making (MCDM) and operations research communities have also worked on multiobjective ....

Parks, G. T., Miller, I. (1998). Selective Breeding in a Multiobjective Genetic Algorithm. In Fifth International Conference on Parallel Problem Solving from Nature (PPSN-V), pages 250--259. Springer.

....leur attribue le rang 2 (fig.12) Ce processus est r eit er e jusqu a ce que tous les individus de la population aient un rang. Cette m ethode de ranking a et e utilis ee dans les AGs pour la r esolution de plusieurs PMO : arbre recouvrant minimum [102] sac a dos [1] conception de r eacteurs [64], distribution de l eau [36] etc. La probabilit e de s election est ensuite affect ee a chaque individu en se basant sur le rang, en appliquant par exemple une m ethode similaire a celle de Baker propos ee pour les probl emes uni objectifs [3] La probabilit e qu un individu de rang n de la ....

....est son rang. Ceci permet d encourager implicitement la diversit e. L elitisme est tr es utilis e dans le processus de s election. Il consiste par exemple a r ealiser la s election des individus aussi bien de la population courante que des solutions non domin ees trouv ees pendant la recherche [64][58] fig.15) L equation VII A est donc modifi ee de la mani ere suivante : pn = N Gamma A N S(N 1 Gamma Rn ) Rn Gamma 2 N (N Gamma 1) o u A est le nombre d individus s electionn es a partir de l ensemble PO courant. B. M ethodes de maintenance de la diversit e Dans la ....

G. T. Parks and I. Miller. Selective breeding in a multiobjective genetic algorithm. In Parallel Problem Solving from Nature PPSN'5, pages 250--259, Amsterdam, Sept 1998. Springer-Verkag.

.... an EMOO technique (see for example [11, 5] The main emphasis has been on using an external le that stores nondominated vectors found during the evolutionary process which are reinserted later in the population (this can be seen as a form of elitism in the context of multiobjective optimization [8, 13, 18]) Following the same line of thought of this current research, we decided to develop an approach in which we would use a GA with a very small population size and a reinitialization process (the so called micro GA) combined with an external le to store nondominated vectors previously found. ....

Georey T. Parks and I. Miller. Selective Breeding in a Multiobjective Genetic Algorithm. In A. E. Eiben, M. Schoenauer, and H.-P. Schwefel, editors, Parallel Problem Solving From Nature | PPSN V, pages 250-259, Amsterdam, Holland, 1998. Springer-Verlag.

.... 1993; Horn et al. 1994; Srinivas and Deb, 1994) Later, these approaches (and variations of them) were successfully applied to various multiobjective optimization problems (Ishibuchi and Murata, 1996; Cunha et al. 1997; Valenzuela Rend on and UrestiCharre, 1997; Fonseca and Fleming, 1998; Parks and Miller, 1998). In recent years, some researchers have investigated particular topics of evolutionary multiobjective search, such as convergence to the Pareto optimal front (Van Veldhuizen and Lamont, 1998a; Rudolph, 1998) niching (Obayashi et al. 1998) and elitism (Parks and Miller, 1998; Obayashi et al. ....

.... and Fleming, 1998; Parks and Miller, 1998) In recent years, some researchers have investigated particular topics of evolutionary multiobjective search, such as convergence to the Pareto optimal front (Van Veldhuizen and Lamont, 1998a; Rudolph, 1998) niching (Obayashi et al. 1998) and elitism (Parks and Miller, 1998; Obayashi et al. 1998) while others have concentrated on developing new evolutionary techniques (Laumanns et al. 1998; Zitzler and Thiele, 1999) For a thorough discussion of evolutionary algorithms for multiobjective optimization, the interested reader is referred to Fonseca and Fleming ....

[Article contains additional citation context not shown here]

Parks, G. T. and Miller, I. (1998). Selective breeding in a multiobjective genetic algorithm. In Eiben, A. E., B ack, T., Schoenauer, M. and Schwefel, H.-P., editors, Fifth International Conference on Parallel Problem Solving from Nature (PPSN-V), pages 250--259, Springer, Berlin, Germany.

.... applicability, and their lack of assumptions about the decision maker and several other methods of assigning fitness based on some form of Pareto ranking have been devised, e.g. 4, 10] More recently, elitism has been shown to improve the performance of multiobjective GAs (for example see [21]) and a very elegant method of exploiting co evolution to perform fitness assignment in an elitist GA was put forward by Zitzler and Thiele [30, 31, 32] The latter has been compared to some of the most popular MOGAs, on a range of problems and test functions, with very positive results. Some ....

G. T. Parks and I. Miller. Selective Breeding in a Multiobjective Genetic Algorithm. In A. E. Eiben, M. Schoenauer, and H.-P. Schwefel, editors, Parallel Problem Solving From Nature --- PPSN V, pages 250--259, Amsterdam, Holland, 1998. Springer-Verlag.

....Solutions from P known (t) are sometimes inserted into the mating population in an attempt to maintain diversity (Todd and Sen, 1997; Ishibuchi and Murata, 1998) These implementations never reduce P known (t) s size except when removing solutions whose evaluated vectors become dominated. Although Parks and Miller (1998) implement an archive of Pareto optimal solutions, solutions in P current (t) are not always archived (placed in P known (t) archiving occurs only if a solution is sufficiently dissimilar from those already resident (clustering) If a new solution is added, any archive members no longer Pareto ....

Parks, G. T. and Miller, I. (1998). Selective Breeding in a Multiobjective Genetic Algorithm. In Eiben, A. E., B ack, T., Schoenauer, M. and Schwefel, H.-P., editors, Parallel Problem Solving from Nature - PPSN V. Springer, Berlin, Germany.

.... and Goldberg 1994; Srinivas and Deb 1994) Later, these approaches (and variations of them) were successfully applied to various multiobjective optimization problems (Ishibuchi and Murata 1996; Cunha, Oliviera, and Covas 1997; Valenzuela Rend on and Uresti Charre 1997; Fonseca and Fleming 1998; Parks and Miller 1998). In recent years, some researchers have investigated particular topics of evolutionary multiobjective search, such as convergence to the Pareto optimal front (Veldhuizen and Lamont 1998a; Rudolph 1998) niching (Obayashi, Takahashi, and Takeguchi 1998) and elitism (Parks and Miller 1998; ....

.... 1998; Parks and Miller 1998) In recent years, some researchers have investigated particular topics of evolutionary multiobjective search, such as convergence to the Pareto optimal front (Veldhuizen and Lamont 1998a; Rudolph 1998) niching (Obayashi, Takahashi, and Takeguchi 1998) and elitism (Parks and Miller 1998; Obayashi, Takahashi, and Takeguchi 1998) while others have concentrated on developing new evolutionary techniques (Laumanns, Rudolph, and Schwefel 1998; Zitzler and Thiele 1999) For a thorough discussion of evolutionary algorithms for multiobjective optimization, the interested reader is ....

[Article contains additional citation context not shown here]

Parks, G. T. and I. Miller (1998). Selective breeding in a multiobjective genetic algorithm. In A. E. Eiben, T. Back, M. Schoenauer, and H.-P. Schwefel (Eds.), Fifth International Conference on Parallel Problem Solving from Nature (PPSN-V), Berlin, Germany, pp. 250-259. Springer.

....paper, we describe the principle of multi objective optimization and then discuss a number of evolutionary algorithms. Since evolutionary algorithms deal with a population of solutions [15] it is logical that they can be used to find multiple Pareto optimal solutions in one single simulation run [3, 11, 18, 24, 1 26, 35]. We describe one such algorithm Non dominated sorting GA or NSGA [32] in somewhat greater details. We present simulation results of NSGA on two problems. Although most research on multi objective evolutionary algorithms have concentrated their efforts in developing new and efficient search ....

Parks, G. T. and Miller, I. (1998). Selective breeding in a multi-objective genetic algorithm. Proceedings of the Parallel Problem Solving from Nature, V, 250--259.

....test functions, multiobjective performance assessment. 1 Introduction Multiobjective optimization using evolutionary algorithms has been investigated by many authors in recent years (Bentley and Wakefield, 1997; Fonseca and Fleming, 1995a; Horn et al. 1994; Horn and Nafpliotis, 1994; Parks and Miller, 1998; Schaffer, 1985; Srinivas and Deb, 1994) However, in some real world optimization problems, the performance of the genetic algorithm is overshadowed by local search methods such as simulated annealing and tabu search, either when a single objective is sought or when multiple objectives have been ....

....solution. We tried both of these, however, and found the results to be very poor. Echoing Horn et al. s findings, we found that the use of a nontrivially sized comparison set is crucial to reasonable results. We note that the idea of maintaining a list of nondominated solutions is not new. Parks and Miller (1998) recently describe a MOGA that also maintains an archive of nondominated solutions. In their case, the overall algorithm is much more complicated than PAES, and the archive is not just used as a repository and a source for comparisons but also plays a key role as a pool of possible parents for ....

[Article contains additional citation context not shown here]

Parks, G. T. and Miller, I. (1998). Selective Breeding in a Multiobjective Genetic Algorithm. In Eiben, A. E., B ack, T., Schoenauer, M., and Schwefel, H.-P., editors, Proceedings of the Fifth International Conference on Parallel Problem Solving from Nature, pages 250--259, Springer, Berlin, Germany.

No context found.

G. T. Parks and I. Miller. Selective Breeding in a Multiobjective Genetic Algorithm. Lecture Notes in Computer Science, 1498:250--259, 1998.

No context found.

Georey T. Parks and I. Miller. Selective breeding in a multiobjective genetic algorithm. In A. E. Eiben, M. Schoenauer, and H.-P. Schwefel, editors, Parallel Problem Solving From Nature | PPSN V, pages 250-259, Amsterdam, Holland, 1998. Springer-Verlag.

No context found.

Parks, G. T. and Miller, I. (1998). Selective breeding in a multi-objective genetic algorithm. Proceedings of the Parallel Problem Solving from Nature, V, 250--259.

No context found.

G. T. Parks and I. Miller. Selective Breeding in a Multiobjective Genetic Algorithm. Lecture Notes in Computer Science, 1498:250-259, 1998.

No context found.

G. T. Parks and I. Miller. Selective breeding in a multiobjective genetic algorithm. In A. E. Eiben et al., editors, Parallel Problem Solving from Nature -- PPSN V, pages 250--259, Berlin. Springer, (1998).

First 50 documents

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC