Ralph E. Steuer. Multiple criteria optimization: theory, computation and application. Wiley, 1986.  Home/Search   Document Not in Database   Summary   Related Articles   Check

.... In domains involving goals with explicit expressions, one could formulate the design problem using an optimization methodology [17] Such single objective formulations have been argued to be constraining for actual design problems [32] Instead, multi objective function formulations could be used [12, 28, 32]. The methodology presented here is compatible with these multi objective function algorithms, in that one can use them to solve the formulations presented here, when the domain has sufficient formalization (performance parameter equations) 28] The focus of this paper is on formally specifying ....

.... function formulations could be used [12, 28, 32] The methodology presented here is compatible with these multi objective function algorithms, in that one can use them to solve the formulations presented here, when the domain has sufficient formalization (performance parameter equations) [28]. The focus of this paper is on formally specifying the multi criteria objective function, not methods for finding its global peak. In the preliminary design domain, the degree of specification of candidate models is usually incomplete. The method of imprecision can still be used, however, to ....

[Article contains additional citation context not shown here]

STEUER,R.Multiple Criteria Optimization: Theory, Computation, and Application. J. Wiley, New York, NY, 1986.

....that the trade o# curves may include non convex parts. The results obtained from the case studies in Section 5 also proves the possibility for the non convexity of the Pareto front. Instead of the weighted sum method, the lexicographic weighted Chebyshev metric method, introduced by Steuer [7], is employed to plot the Pareto front of each subproblem P (N ) This method makes it possible to compute the whole Pareto front [7] regardless of its shape, by minimizing the distance to a reference point. The reference point selection is based on the ideal point, z where z i = min x#S f i ....

....possibility for the non convexity of the Pareto front. Instead of the weighted sum method, the lexicographic weighted Chebyshev metric method, introduced by Steuer [7] is employed to plot the Pareto front of each subproblem P (N ) This method makes it possible to compute the whole Pareto front [7] regardless of its shape, by minimizing the distance to a reference point. The reference point selection is based on the ideal point, z where z i = min x#S f i (v) Then the reference point is given by i = z # i where # i 0 for all i = 1, Q. In this study, each # i is selected ....

R. E. Steuer, Multiple Criteria Optimization: Theory, Computation and Application, John Wiley and Sons, New York, 1985.

....3. 1.000 15. 0.957 7, 1.000 12. 1.000 7 Lossy Dielectric Materials ( lr=l. j0. e(1 GHz) 1 GHz) e(f) f e=er je, er(f) f e(1 GHz) g e(l GHz) 5. 0.861 8. 0.569 8, 0.778 10. 0.682 10. 0.778 6, 0. 861 Relaxation type Magnetic Materials #]2 #ff (fandf in GHz) e r = 15 jO 9 35. 0.8 10 35. 0.5 11 30. 1.0 12 18. 0.5 13 20. 1.5 2.5 14 30. 15 30. 16 25. 2.0 3.5 variety in the behavior of the multilayer, but not necessarily in its physical construction. III. NUMERICAL PESTJETS In this section, we present the results of applying the aforementioned Pareto GA ....

....15. 0.957 7, 1.000 12. 1.000 7 Lossy Dielectric Materials ( lr=l. j0. e(1 GHz) 1 GHz) e(f) f e=er je, er(f) f e(1 GHz) g e(l GHz) 5. 0.861 8. 0.569 8, 0.778 10. 0.682 10. 0.778 6, 0.861 Relaxation type Magnetic Materials #]2 #ff (fandf in GHz) e r = 15 jO 9 35. 0. 8 10 35. 0.5 11 30. 1.0 12 18. 0.5 13 20. 1.5 2.5 14 30. 15 30. 16 25. 2.0 3.5 variety in the behavior of the multilayer, but not necessarily in its physical construction. III. NUMERICAL PESTJETS In this section, we present the results of applying the aforementioned Pareto GA algorithms as ....

[Article contains additional citation context not shown here]

R. E. Steuer, Multiple Criteria Optimization: Theory, Computation, Application. Malabar, FL: Krieger, 1989.

....weight W(x) is assigned to auction to scale the risk value according to the probability of having risk. 4. 2 Heuristic search for best gain and risk Since there are two goals (safety and profitability) to optimize, the core of the Gain Risk model is a multiple criteria optimization problem [Steuer, 1986]. One possible solution is to use multiple objective linear programming (MOLP) The alternative solutions are classical AI search techniques, such as A or beam search. The UMBCTAC agent runs a relatively simple heuristic search which has two stages. In the first stage, we prune those ....

Steuer, R. Multiple criteria optimization: Theory, Computation and Application. Chichester, John Wiley & Sons: New York.

....algorithm for one gives an exact algorithm that solves the other. For instance, if we only need a solution for a xed k or a xed c, using binary search we can give an answer to one problem using a routine for the other. In fact, the whole feasible region can be described by the Pareto points [Ste86] that is, the undominated pairs of k and c that induce feasible colorings. For the case of approximations, that is not true anymore. Having an approximation algorithm for one problem does not imply the same for the other. This situation is evident for this pair of problems because one ....

R. E. Steuer. Multiple Criteria Optimization: Theory, Computation and Application. Wiley, New York, 1986.

....Figure 1. Illustration of the objective functions on an edge. notation generalizes to the nondominated set Z Par de ned below. Solving the Q criteria semi obnoxious network location problem means nding the set of ecient points. For an introduction to multiple criteria analysis see Steuer [12]. The de nition of eciency is as follows. De nition 1. A solution x 2 G(V; E) to (4) is ecient (Pareto optimal) i there does not exist another solution x 2 G(V; E) to (4) such that f ( x) x) 8q 2 Q and 9q 2 Q s:t: f ( x) f (x) Otherwise x is inecient. The set of all ....

R.E. Steuer. Multiple criteria optimization: Theory, Computation, and Application. Wiley, New York, 1986.

....problem could be solved. 4.4 Formulating the objective As most optimization problems are multiobjective in nature, there are many methods available to tackle this kind of problems. References to multiobjective optimization could be found in Hwang et al. 40] Ringuest [69] and Steuer [82] and with applications to engineering design in Eschenauer et al. 22] and Osyczka [61] Generally, the multiobjective optimization problem (MOOP) can be handled in four different ways depending on when the decision maker articulates his preference concerning the different objectives: never, ....

....to prototype driven. A more thorough discussion of these methods is presented in Paper [XIII] 4.4. 1 No preference articulation There are methods that do not use any preference information, e.g. the Min Max formulation and global criterion method, see Hwang et al. 40] Osyczka [61] and Steuer [82]. The Min Max formulation is based on minimization of the relative distance from a candidate solution to the utopian solution F , see Figure 7. The distance between a solution vector and the utopian vector is typically expressed as a v p norm. 4.4.2 Priori articulation of preference ....

[Article contains additional citation context not shown here]

STEUER R., Multiple criteria optimization: theory, computation and application, John Wiley & Sons, Inc, New York, 1986.

....has been significant interest in using genetic programming for problems with multiple objectives [2] In general these solvers are aimed for small problems. GAMS and the solvers supported by GAMS does not have facilities for these solution procedures. More information can be found in for instance [12], 1] 2. Goal programming A special form of multiple objective programming is goal programming[4] Goal programming deals with multiple, possibly conflicting goals, which may be measured in nonhomogeneous units. An objective f i (x) is reformulated into a goal by considering an aspiration level ....

Ralph E. Steuer, Multiple criteria optimization : Theory, computation, and application, Wiley, 1986.

.... Such applications include neural network learning, digital signal and image processing, structural optimization, engineering design, computer aided design (CAD) for VLSI, database design and processing, nuclear power plant design and operation, mechanical design, and chemical process control [68, 145, 180]. Due to the availability of a lot of unconstrained optimization algorithms, many real applications that are inherently nonlinear and constrained have been solved in various unconstrained forms. Optimal or good solutions to these applications have significant impacts on system performance, such as ....

R. E. Steuer. Multiple Criteria Optimization: Theory, Computation and Application. Krieger Publishing Company, 1989.

..... 131 vii List of Tables viii List of Figures ix Introduction Many engineering applications can be formulated as constrained nonlinear programming problems (NLPs) Examples include production planning, computer integrated manufacturing, chemical control processing, and structure optimization [65, 121, 148]. 1.1 Problem Definition Constrained NLPs can be solved by existing methods if they are specified in well defined formulae that are di#erentiable and continuous. However, only special cases can be solved when they do not satisfy the required assumptions. For instance, sequential quadratic ....

R. E. Steuer. Multiple Criteria Optimization: Theory, Computation and Application. Krieger Publishing Company, 1989.

....and constraint functions are nonlinear. Typical applications include signal processing, structural optimization, neuralnetwork design, VLSI design, database design and processing, nuclear power plant design and operation, mechanical engineering, physical sciences, and chemical process control [144, 193, 17, 47, 56]. A general goal in solving nonlinear constrained optimization problems is to find feasible solutions that satisfy all the constraints. This is not an easy task because nonlinear constrained optimization problems are normally NP hard [63] In practice, the di#culties in solving a nonlinear ....

R. E. Steuer. Multiple Criteria Optimization: Theory, Computation and Application. Krieger Publishing Company, 1989.

....points (and corresponding solutions) such that 3z,z: A such that z z: i.e. set A is composed of mutually nondominated points. The point z composed of the best attainable objective function values is called the ideal point: z = max f (x)lx D j = 1 . J. Range equalization factors ([26], ch. 8.4.2) are defined in the following way: 7r = j=l . J R where R i is the (approximate) range of objective zi in the nondominated set, or D or A. Objective function values multiplied by range equalization factors are called normalized objective function values. 2 Weighted linear ....

....if the optimum is not unique then some of the optima may be dominated, but must have at least one objective component equal to a Pareto optimal solution. For each Pareto optimal solution x there exists a weighted Tchebycheff scalarizing function s such that x is a global optimum (minimum) of s ([26], ch. 14.8) Weight vectors that meet the following conditions: J j=l are called normalized weight vectors. Minimization of the weighted Tchebycheff scalarizing function corresponds to a min max problem. The problem can be, however, transformed to the following one: minimize c (P2) s.t. ....

Steuer R.E. (1986), Multiple Criteria Optimization - Theory, Computation and Application, Wiley, New York.

....simultaneously all the objectives, selection of the best compromise solution requires taking into account preferences of the DM. Under very weak and generally accepted assumptions about the DM s preferences the best compromise solution belongs to the set of, so called, Pareto optimal solutions ([154], ch. 6.6 6.7) Research activities within the area of multiple objective optimization traditionally concentrated on multiple objective linear programming (MOLP) see e.g. 154] 181] In recent years, the demand for new applications and the increasing power of computers resulted in growing ....

....about the DM s preferences the best compromise solution belongs to the set of, so called, Pareto optimal solutions ( 154] ch. 6.6 6. 7) Research activities within the area of multiple objective optimization traditionally concentrated on multiple objective linear programming (MOLP) see e.g. [154], 181] In recent years, the demand for new applications and the increasing power of computers resulted in growing interest in computationally hard MOO problems, e.g. non linear and combinatorial optimization problems. This work focuses on multiple objective combinatorial optimization (MOCO) ....

[Article contains additional citation context not shown here]

Steuer R.E. (1986), Multiple Criteria Optimization - Theory, Computation and Application, Wiley, New York.

....required from the decision maker are presented with weights, targets, trade offs and goal levels to formulate the problem. Steuer proposed the objective function of a linear goal programming to be a weighting representation of second objective functions with the sum of these weights equal to unity [17]. Allowing these weights to vary within the range between 0 and 1 a decision maker performs the sensitivity analysis of all these weights simultaneously. The difficulty with the Steueris weight technique is that in many situations a decision maker is unwilling to specify the weights [14] Also, ....

Steuer, R. (1986), Multiple criteria optimization: Theory, Computation and Application. Chichester, John Wiley & Sons: New York.

....if and only if there is no other x # X such that # # (x) # # # (x ) for all # =1###### and the inequality is strict for at least one #. In the literature on multiobjective programming the corresponding criterion vector (# # (x ) #### # (x ) # IR is called nondominated solution [25, 30]. Depending on the assumptions made about the DMs value functions and the decision set it is possible to derive di erent necessary and sucient conditions for Pareto optimality. These conditions are usually more appealing from the computational point of view than the original de nition. In the ....

R.E. Steuer, Multiple Criteria Optimization: Theory, Computation and Application (John Wiley, New York, 1986).

.... them, to our knowledge, is complete (i.e. can provide the whole nondominated frontier) and the usual way to tackle with it is to split the multi objective problem into many COPs (each with one function to be optimized) and solve them independently, or to combine the various functions in only one [9]. In other problems, the partial order comes directly from the formalization. We considered a problem of 3D object recognition based on CSP solving, and found out that the solution tuples had to be ranked in a partial order. In the following of the paper, we define the Partially ordered ....

....is a totally ordered set. 1 q Fig. 3. Ratio of the geometric mean of the timing results for the B B Splitting B B PCOP We consider as solution of the MOP the set of non dominated CSP solutions. The solutions are sorted through the ordering defined in T m (also called the criterion space [9]) Definition 6. Given A, B # T m , we say that A # p B i# # m i=1 A i # B i . # p is a partial ordering on the set of possible result T m , moreover, it is a lattice. The mapping m of the PCOP coincides with the vector function f of the MOP. The constraints added in the B B ....

[Article contains additional citation context not shown here]

R.E. Steuer. Multiple Criteria Optimization: Theory, Computation, and Application. Wiley, New York, 1986.

....the original MSM T and the values ae 0:5 are symmetric, with the roles of the training and tuning sets reversed. 5. 2 Relationship to Pareto optimality Considering the tuning and training errors separately in the objective function suggests the field of multi objective mathematical programming [83]. This area 72 of optimization considers problems of the form vector min x2S 2 6 6 6 6 6 6 6 4 f 1 (x) f 2 (x) f k (x) 3 7 7 7 7 7 7 7 5 (5.3) Here, multiple, possibly competing objectives are to be optimized. The possible solutions to such a problem are known as ....

R. E. Steuer. Multiple Criteria Optimization: Theory, Computation, and Application. John Wiley and Sons, 1986.

....for sustainable decision making. Depending on the set of alternatives, Multiple Criteria Decision Making (MCDM) can involve a choice problem (when there is a small, explicit list of alternatives) or a design problem (when an infinite set of alternatives is implicitly defined by constraints) (Steuer 1986). The different approaches to Multiple Criteria Decision Making (MCDM) can be classified according to the time when the decision maker s preferences enter the formal decision making process (Hwang and Masud 1979) no articulation of preference information, a priori articulation of ....

Steuer, R. E. (1986). Multiple Criteria Optimization: Theory, Computation, and Application. New York: Wiley.

....ideal objective vector. The components z ideal i are obtained by minimizing each of the objective functions individually. Other functions than the weighted sum are used, because the weakness of weighting methods is that not all Pareto optimal solutions can be found unless the problem is convex (Steuer 1986). Weighted sums in conjunction with multi objective tabu search do not seem to be a practical problem, since every iteration a new sum is calculated. Above adaptations contain weights, which may influence or even steer the search process. Thus, weights must be set carefully. Different weight ....

Steuer, R. E. (1986). Multiple Criteria Optimization: Theory, Computation, and Application. John Wiley & Sons.

....fast non dominated sort(R t ) P t 1 = and i = 1 until jP t 1 j jF i j N P t 1 = P t 1 [ F i crowding distance assignment(F i ) i = i 1 Sort(F i ; n ) P t 1 = P t 1 [ F i [1 : N jP t 1 j) Q t 1 = make new pop(P t 1) t = t 1 Fig. 1. NSGA II algorithm is shown. non domination [10] and di erent non dominated fronts F 1 , F 2 , and so on are found. The algorithm is illustrated in the following: The new parent population P t 1 is formed by adding solutions from the rst front F 1 and continuing to other fronts successively till the size exceeds N . Individuals of each front ....

Steuer, R. E. (1986). Multiple criteria optimization: Theory, computation, and application. New York: Wiley.

....However, a form of adaptive discretization might able to alleviate the inherent inefficiency of such a method. 1.2. 3 Total Linearization Linearizing all functions involved results in a linear multicriteria optimization problem, for which Simplex based solver strategies exist (see, e.g. [98, 27, 87, 88]) In singlecriteria optimization, a complete linearization is seldom used, since the resulting linear problem is rather large. From a theoretical point of view, Simplex based methods of todays standards still lack the efficiency of other methods, especially interior point approaches. Moreover, ....

....criteria, i.e. an ordering of importance of the components of the objective function vector. In this case, the ordering has to be specified, which is in itself a set of parameters. In other parameter free techniques, the optimization process is augmented by an interactive procedure (see, e.g. [87, 65]) adding an additional burden to the task of the decision maker. This particular approach is infeasible for real time optimization and problematic in large scale optimization. Moreover, usually only a small amount of interaction ever takes place in practice, although most methods require a rather ....

[Article contains additional citation context not shown here]

R. E. Steuer. Multiple Criteria Optimization: Theory, Computations, and Application. Wiley, New York, 1986.

....required from the decision maker are presented with weights, targets, trade offs and goal levels to formulate the problem. Steuer proposed the objective function of a linear goal programming to be a weighting representation of second objective functions with the sum of these weights equal to unity (Steuer 1986). Allowing these weights to vary within the range between 0 and 1 a decision maker performs the sensitivity analysis of all these weights simultaneously. The difficulty with the Steuer s weight technique is that in many situations a decision maker is unwilling to specify the weights (Lootsma ....

Steuer, R. (1986). Multiple criteria optimization: Theory, Computation and Application. Chichester, John Wiley & Sons: NewYork.

..... 111 xv Chapter 1 Introduction Many engineering applications can be formulated as constrained nonlinear programming problems (NLPs) Examples include production planning, computer integrated manufacturing, chemical control processing, and structure optimization [65, 121, 148]. 1.1 Problem Definition Constrained NLPs can be solved by existing methods if they are specified in well defined formulae that are differentiable and continuous. However, only special cases can be solved when they do not satisfy the required assumptions. For instance, sequential quadratic ....

R. E. Steuer. Multiple Criteria Optimization: Theory, Computation and Application. Krieger Publishing Company, 1989.

....the weighted mean [ i E (A ij ) 1 i )E (A ij ) b i i = 1 i )E (B i ) i E (B i ) and c ij = F(C ij ) Remark that we allow di erent satisfaction requirements for the di erent constraints. 1 MOFAC method 15 1. 4 Interactive procedure In the literature (see, e.g. Steuer 1985, Vanderpooten and Vincke 1989, S lowi nski 1997) the authors proposed various iterative methods to determine ecient solutions of a MOLP problem. They consist of successive unicriterion optimization of a weighted aggregated objective. At each step, new ecient solutions are computed and in general ....

Steuer, R. (1985). Multiple Criteria Optimization: Theory, Computation and Applications, John Wiley and Sons.

....by such a procedure. This is a characteristic of multicriteria optimization problems with a nonconvex domain (e.g. involving boolean variables) and requires a special treatment. For some problems (e.g. assignment and knapsack) specific procedures have been developed to catch them (see for example [33] and [34] for more details) No procedure seems to be designed for multicriteria jobshop scheduling. The development of such algorithms is far beyond the scope of this paper and is independent of our problem: the study of jobshop scheduling with imprecise durations. Nevertheless, it is worthwhile ....

R. Steuer, Multiple Criteria Optimization: Theory, Computation and Applications. New York: John Wiley & Sons, 1985.

....this paper we suggest to use cones and norms, two concepts well known in convex analysis, to construct piecewise linear approximations of the nondominated set of convex multicriteria programming problems. Both cones and norms have been used in multicriteria programming quite extensively (see e.g. Steuer (1986) and Kaliszewski (1987) but, to our knowledge, Kaliszewski (1994) is the only other source to combine both concepts in order to describe and solve multicriteria programs. There have been quite a few approximation approaches developed for bicriteria convex as well as general problems, see, e.g. ....

STEUER, R.E. (1986): Multiple Criteria Optimization: Theory, Computation, and Application. Wiley, New York.

....objective program, it may feature globally as well as locally ecient solutions that can be found by means of some suitable scalarizations specially developed to handle non convexity. All the globally ecient solutions can be found by means of the lexicographic weighted Tchebyche approach (see [27]) while the locally ecient solutions can be generated using the augmented Lagrangian approach (see [28] In order to avoid treating (4) in this general methodological framework and to obtain speci c and more e ective approaches, we focus on the special case of line barriers with passages but ....

R.E. Steuer. Multiple Criteria Optimization - Theory, Computation and Application. John Wiley, New York, 1986.

....point. Di erent norms have been used to measure the distance between the solutions and the utopia point. In particular, the family of L p norms has been extensively studied by many researchers, including [Yu, 1973] Zeleny, 1973] Gearhart, 1979] Wierzbicki, 1980] Steuer and Choo, 1983] [Steuer, 1986], and many others. The l 1 norm and the augmented l 1 norm turned out to be very useful in generating nondominated solutions of general continuous or discrete multiple criteria programs and led to the well known weighted (augmented) Tchebyche scalarization and its variations. Kaliszewski, 1987] ....

....norm so that this point is a unique minimizer of the related block norm scalarization. In the proof, to show the existence of the desired block norm we use the l 1 norm, and thus not an oblique norm. The result gives another interpretation of the results on the weighted Tchebyche approach in [Steuer, 1986] and illustrates the idea of introducing block norms to multiple criteria programming. Theorem 3.1 Let z 2 N . Then there exists a block norm so that z uniquely minimizes min z2Z (z) min x2S (f(x) Proof. Recall that we assumed without loss of generality z = 0. De ne the unit ....

Steuer, Ralph E. (1986). Multiple Criteria Optimization: Theory, Computation, and Application. Wiley, New York.

....It was then proposed to further use this utility function to explore the efficient solutions in a neighborhood of the candidate solution. In Chen et al. 1998, the advantages of the CP approach over the WS method in locating the efficient multiobjective robust design solutions (Pareto points) (Steuer, 1986) are thoroughly illustrated both in principle and through the example problems. The derivation of a family of quadratic utility functions at the considered efficient solution is demonstrated, however, the methodology on how to choose the parameter a that yields the most accurate quadratic function ....

.... is provided by Das and Dennis (1997) Though the weights representing relative importance are used as the preference structure when applying CP, it has been mathematically proven that CP is superior to the weighted sum (WS) method in locating the efficient solutions, or the so called Pareto points (Steuer, 1986). Due to Geoffrion (1968) for every Pareto point of a convex multiobjective optimization problem there exists a (nonzero) vector weight w 0 such that this Pareto point is an optimal solution of the WSP(w) However, not every Pareto solution of a general (nonconvex) problem can be found by ....

Steuer, R. E., 1986, Multiple Criteria Optimization: Theory, Computation and Application, John Wiley, New York .

.... point (Zeleny s axiom of choice) Though the weights representing relative importance are used as the preference structure when applying CP, it has been mathematically proven that CP is superior to the weighted sum (WS) method in locating the efficient solutions, or the so called Pareto points (Steuer, 1986). However, there are few applications of CP to mechanical engineering design problems. Miura and Chargin (1996) develop a variation of CP and apply it to optimal structural design. Athan and Papalambros (1996) do not refer to CP but propose to minimize the sum of the exponentially weighted ....

....developed Compromise Programming (CP) an approach based on a procedure that finds an efficient point closest to the utopia point. Since to measure the distance between an efficient point and the utopia point one may use different metrics (mathematical measures of distance between points, see Steuer, 1986), the general compromise programming problem is formulated as: minimize f(x) u subject to x X, 2.6) where . denotes the metric of choice. For a weighted L p metric, the distance between two points r, s in R m is given by: r s w p = w i r i s i i =1 m ) p ) 1 p , ....

Steuer, R. E., 1986, Multiple Criteria Optimization: Theory, Computation and Application, John Wiley, New York .

....= 1; M and M 1) any two solutions x (1) and x (2) having P decision variables each) can have one of two possibilities one dominates the other or none dominates the other. A solution x (1) is said to dominate the other solution x (2) if both the following conditions are true [22]: 1. The solution x (1) is no worse (say the operator denotes worse and denotes better) than x (2) in all objectives, or f j (x (1) f j (x (2) for all j = 1; 2; M objectives. 2. The solution x (1) is strictly better than x (2) in at least one objective, or f j (x ....

....j) Q t 1 = make new pop(P t 1 ) t = t 1 First, a combined population R t = P t [ Q t is formed. This allows parent solutions to be compared with the child population, thereby ensuring elitism. The population R t is of size 2N . Then, the population R t is sorted according to non domination [22]. The sorting procedure classi es the population into several non dominated fronts, F 1 , F 2 , and so on. The new parent population P t 1 is formed by adding solutions from the rst front F 1 and continuing to other fronts successively till the size exceeds N . Thereafter, the solutions of the ....

Steuer, R. E. (1986). Multiple criteria optimization: Theory, computation, and application. New York: Wiley.

....fast non dominated sort(R t ) P t 1 = and i = 1 until jP t 1 j jF i j N P t 1 = P t 1 [ F i crowding distance assignment(F i ) i = i 1 Sort(F i ; n ) P t 1 = P t 1 [ F i [1 : N jP t 1 j) Q t 1 = make new pop(P t 1) t = t 1 Fig. 1. NSGA II algorithm is shown. non domination [10] and di erent non dominated fronts F 1 , F 2 , and so on are found. The algorithm is illustrated in the following: The new parent population P t 1 is formed by adding solutions from the rst front F 1 and continuing to other fronts successively till the size exceeds N . Individuals of each front ....

Steuer, R. E. (1986). Multiple criteria optimization: Theory, computation, and application. New York: Wiley.

....op erationnelle, et la facilit e relative de traiter de tels probl emes, et d autre part l abondance des cas pratiques qui peuvent etre formul es sous forme lin eaire. Ainsi, un certain nombre de logiciels ont vu le jour depuis le d eveloppement de la m ethode du simplexe multi objectif [101][80]. Nous nous int eressons dans cet article, principalement, aux probl emes d optimisation combinatoire multi crit ere pour lesquels relativement peu de travaux ont et e r ealis es. Dans cette section, nous pr esentons quelques exemples aussi bien acad emiques qu industriels. 5 f2 f1 9 5 2 ....

....pour la r esolution de probl emes d optimisation combinatoire multi crit eres. Une analyse critique de chaque classe de m ethodes est aussi r ealis ee. Dans le pass e, la plupart des articles de synth ese sur les m ethodes d optimisation multi objectif concernaient la programmation math ematique [80][7] Des articles de synth ese sur l application de m etaheuristiques ont vu r ecemment le jour, mais portent sur des m ethodes sp ecifiques comme les algorithmes evolutionnaires [9] 25] 16] qui ont recu un int eret croissant ces derni eres ann ees. D autres m etaheuristiques ont et e utilis ees ....

R. Steuer. Multiple criteria optimization: Theory, computation and application. Wiley, New York, 1986.

....A definition of the problem of BS positioning is given in a multiobjective optimization context. The model deals with specific constraints due to the engineering of cellular radio network. Many search algorithms have been used to solve multiobjective combinatorial optimization problems [Tal00][Ste86]. Exact algorithms such as branch and bound [SK99] and dynamic programming [CMM90] have been used to solve small instances of bi objective problems. The design problem is a complex multiobjective combinatorial problem, where a heuristic approach is required. Some metaheuristics have been ....

R. Steuer. Multiple criteria optimization: Theory, computation and application. Wiley, New York, 1986.

No context found.

Ralph E. Steuer. Multiple criteria optimization: theory, computation and application. Wiley, 1986.

No context found.

Steuer Ralph E. Multiple Criteria Optimization: Theory, Computation and Application. Wiley, 1986.

No context found.

Steuer Ralph E. (1986). Multiple Criteria Optimization: Theory, Computation and Application. Wiley.

No context found.

R.E. Steuer (1985). Multiple Criteria Optimization: Theory, Computation and Application. Wiley, NewYork.

No context found.

Steuer, R., E. Multiple Criteria Optimization - Theory, Computation and Application, Wiley, 1986.

No context found.

Steuer, R. E. (1986). Multiple Criteria Optimization: Theory, Computation, and Application. John Wiley & Sons, New York.

No context found.

R. E. Steuer, Multiple Criteria Optimization: Theory, Comutation and Applicaion, Wiley, New York, 1986.

No context found.

Steuer R., Multiple criteria optimization: theory, computation and application. New York, John Wiley & Sons, Inc., 1986.

No context found.

Steuer, R. E. (1986). Multiple criteria optimization: Theory, computation, and application. New York: Wiley.

No context found.

R. E. Steuer, Multiple Criteria Optimization: Theory, Computation and Application, Wiley, ]986.

No context found.

Steuer, R. (1986), Multiple criteria optimization: Theory, Computation and Application. Chichester, John Wiley & Sons: New York.

No context found.

Steuer, R., Multiple Criteria Optimization: Theory, Computation, and Application , New York: Wiley, 1986.

No context found.

Steuer RE (1986). Multiple Criteria Optimization: Theory, Computation, and Application. Wiley: New York.

No context found.

R.E. Steuer, Multiple Criteria Optimization: Theory, Computation, and Application, Wiley, New York, 1986.

No context found.

Steuer R.E. Multiple Criteria Optimization - Theory, Computation and Application, Wiley, New York, 1986.

No context found.

Steuer, R.E., Multiple Criteria Optimization: Theory, Computation, and Application, John Wiley & Sons, New York, 1986

First 50 documents  Next 50

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