C. A. Coello Coello. Constraint-handling using an evolutionary multiobjective optimization technique. Civil Engineering and Environmental Systems, 17:319--346, 2000.  Home/Search   Document Details and Download   Summary   Related Articles   Check

No context found.

C. A. Coello Coello. Constraint-handling using an evolutionary multiobjective optimization technique. Civil Engineering and Environmental Systems, 17:319--346, 2000.

....of this approach is that the number of subpopulations required increases linearly with the number of constraints of the problem. This has some obvious scalability problems. Furthermore, it is not clear how to determine appropriate sizes for each of the subpopulations used. 3. 3 MOGA Coello [6] proposed the use of Pareto dominance selection to handle constraints in EAs. This is an application of Fonseca and Fleming s Pareto ranking process [11] called Multi Objective Genetic Algorithm, or MOGA) to constraint handling. In this approach, feasible individuals are always ranked higher than ....

....avoids the usual empirical fine tuning of the main genetic operators. Coello s approach uses a real coded GA with universal stochastic sampling selection (to reduce the selection pressure caused by the Pareto ranking process) This approach has been used to solve some engineering design problems [6] in which it produced very good results. Furthermore, the approach showed great robustness and required a relatively low number of fitness function evaluations with respect to traditional penalty functions. Additionally, it does not require any extra parameters. Its main drawback is the ....

[Article contains additional citation context not shown here]

Carlos A. Coello Coello. Constraint-handling using an evolutionary multiobjective optimization technique. Civil Engineering and Environmental Systems, 17:319--346, 2000.

....In Section 3, the most popular multiobjective based constraint handling techniques are discussed. Section 4 presents a comparative study in which four of the techniques discussed in the previous section are tested on several benchmark problems taken from the standard constraint handling literature [31, 11]. Section 5 discusses the results obtained, and Section 6 provides some conclusions and possible paths for future research. 2 Basic Concepts We are interested in the general non linear programming problem in which we want to: Find x which optimizes f( x) 1) g i ( x) 0; i = 1; n (2) ....

....required increases linearly with the number of constraints of the problem. This has some obvious scalability problems when dealing with highly constrained search spaces. Furthermore, it is not clear how to determine appropriate sizes for each of the subpopulations used. 3. 8 MOGA Coello [11] proposed the use of Pareto dominance selection to handle constraints in a genetic algorithm. This is an application of Fonseca and Fleming s Pareto ranking process [18] called Multi Objective Genetic Algorithm, or MOGA) to constraint handling. In this approach, feasible individuals are always ....

[Article contains additional citation context not shown here]

Carlos A. Coello Coello. Constraint-handling using an evolutionary multiobjective optimization technique. Civil Engineering and Environmental Systems, 17:319-346, 2000.

....optimum. At that point the direct use of the objective function will help the EA to approach the optimum, but since some infeasible solutions will still be present in the population, those individuals will be responsible to keep the diversity required to avoid stagnation. More recently, Coello [23] proposed another approach based on nondominance. In this case, tness is assigned to an individual using the following algorithm: Let the vector x i (i = 1; pop size) be an individual in the current population whose size is pop size. The proposed algorithm is the following: 22 To ....

....universal sampling is used) Values produced by fitness( x i ) must be normalized to ensure that the rank of feasible individuals is always higher than the rank of infeasible ones. This approach uses a real coded GA with a simple self adaptive mechanism for crossover and mutation (see [23] for details) and it does not require any additional parameters to maintain diversity in the population (as is normally the case of evolutionary multiobjective optimization techniques [21] Ray et al. 140] proposed an approach in which solutions are ranked separately based on the value of their ....

[Article contains additional citation context not shown here]

Carlos A. Coello Coello. Constraint-handling using an evolutionary multiobjective optimization technique. Civil Engineering and Environmental Systems, 17:319-346, 2000.

.... If both solutions are infeasible, the one with the lowest sum of constraint violation is preferred. 4 Experiments and Results To evaluate the performance of the techniques selected, we decided to use the wellknown benchmark proposed in [23] plus four engineering design problems used in [14]. The full description of the seventeen test functions is the following: 5 5 5 (6) subject to: 505 (7) ....

Carlos A. Coello Coello. Constraint-handling using an evolutionary multiobjective optimization technique. Civil Engineering and Environmental Systems, 17:319--346, 2000.

No context found.

C. A. Coello Coello. Constraint-handling using an evolutionary multiobjective optimization technique. Civil Engineering Systems, Gordon and Breach Science Publishers, 2000.

No context found.

Carlos A. Coello Coello. Constraint-handling using an evolutionary multiobjective optimization technique. Civil Engineering and Environmental Systems, 17, 319--346, (2000).

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