F. Herrera, M. Lozano, and J.L. Verdegay, "A learning process for fuzzy control rules using genetic algorithms.," Fuzzy Sets and Systems, vol. 100, pp. 143--158, 1998.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....avoidance behavior for a mobile robot. The fuzzy controller maps the input perceived by a set of sonar senors to a control action, namely the turn rate of the robot. The majority of applications in the domain of genetic fuzzy systems is concerned with the optimization of fuzzy logic controllers [12], 13] 9] 14] 11] Several authors proposed evolutionary algorithms for learning robotic behaviors implemented by fuzzy control rules [15] 16] 17] 18] The ELF (evolutionary learning of fuzzy rules) system employs a credit assignment mechanism similar to reinforcement learning ....

....to demonstrate a desired behavior. Genetic fuzzy systems have been successfully applied to control system design, modeling, decision making, optimization, classi cation and information retrieval [21] The majority of publications is concerned with fuzzy rule based control system design [15] [12], 22] 9] 10] 14] 11] For two reasons, a fuzzy representation is particularly useful for evolutionary optimization compared to other possible parameterizations of a controller. For many real world problems a mathematical precise and complete solution is not only unnecessary, but often ....

F. Herrera, M. Lozano, and J.L. Verdegay, \A learning process for fuzzy control rules using genetic algorithms.," Fuzzy Sets and Systems,vol. 100, pp. 143-158, 1998.

....avoidance behavior for a mobile robot. The fuzzy controller maps the input perceived by a set of sonar senors to a control action, namely the turn rate of the robot. The majority of applications in the domain of genetic fuzzy systems is concerned with the optimization of fuzzy logic controllers [12], 13] 9] 14] 11] Several authors proposed evolutionary algorithms for learning robotic behaviors implemented by fuzzy control rules [15] 16] 17] PROCEEDINGS OF THE IEEE, VOL. XX, NO. Y, MONTH 2000 2 [18] The ELF (evolutionary learning of fuzzy rules) system employs a credit ....

....to demonstrate a desired behavior. Genetic fuzzy systems have been successfully applied to control system design, modeling, decision making, optimization, classi cation and information retrieval [21] The majority of publications is concerned with fuzzy rule based control system design [15] [12], 22] 9] 10] 14] 11] For two reasons, a fuzzy representation is particularly useful for evolutionary optimization compared to other possible parameterizations of a controller. For many real world problems a mathematical precise and complete solution is not only unnecessary, but often ....

F. Herrera, M. Lozano, and J.L. Verdegay, \A learning process for fuzzy control rules using genetic algorithms.," Fuzzy Sets and Systems, vol. 100, pp. 143-158, 1998.

....(FCAC) uses the mean and peak bit rates and mean burst length to describe the traffic behaviour of each of the multiplexed connections on a node to node basis. The fuzzy rule base in FCAC is automatically designed using a method of learning from examples based on the work done by Herrera et al. [12]. This learning method has been explained in detail in previous studies (see also Ramalho et al. 13] and [14] and it allows to define (a) the fuzzy sets for the fuzzy variables in the antecedent and consequent of each fuzzy rule and (b) a finite set of fuzzy rules able to reproduce the ....

Herrera F., Lozano M., Verdegay J., "A learning process for Fuzzy Control rules using Genetic Algorithms", Dept. of Computer Science and Artificial Intelligence, Univ. Granada, Spain, Technical Report #DECSAI-95108, February 1995.

....the completeness and consistency of the nal KB generated. Since the same antecedent combination can be generated for several rules, the method implicitly makes more exible the rule structure allowing it to have several consequents associated. The second stage, the simpli cation one (proposed in [18]) combines rules and eliminates redundant rules, selecting the most cooperative set of them. It is based on a binary coded GA (one bit for each rule belonging to the previous RB) where each gene indicates the consideration or not of the corresponding rule to belong to the nal RB. Appropriate ....

....in importance, are obtained in each combination considering a speci c generation process. In this contribution, the generation process based on the WM method will be considered. Then, after decomposing each double consequent rule into two independent simple ones, the selection process proposed in [18] and previously described in the M L method is employed to select a subset of the rules best cooperating. The proposal of Nozaki, Ishibuchi, and Tanaka (NIT method) in [25] also employs two consequents in each combination of antecedents to improve the performance. In this case, moreover of ....

[Article contains additional citation context not shown here]

F. Herrera, M. Lozano, J.L. Verdegay, A learning process for fuzzy control rules using genetic algorithms, Fuzzy Sets and Systems 100 (1998) 143-158.

....Systems In [8] a multistage genetic learning process for FRBCSs is proposed, divided into three stages: 1. A fuzzy rule generation process, that obtains a linguistic RB which represents the knowledge extracted from the training samples and veri es the completeness and k consistency properties [13, 14]. 2. A genetic multiselection process that generates di erent KBs. In this process a selection of a rule subset is carried out as well as a learning of a linguistic modi er set, considering the FRM used in the classi cation stage. 3. A genetic tuning process that leads to obtain the best ....

F. Herrera, M. Lozano, and J. L. Verdegay. A learning process for fuzzy control rules using genetic algorithms. Fuzzy Sets and Systems, 100:143-158, 1998.

.... The University of Texas at Austin, 394] Tulane University, 52] Universidad de Granada, 183, 243, 244, 377] University of Bristol, 251] University of California at Berkley, 19] University of Cambridge, 72] University of Durham, 396] University of Edinburgh, 42] University of Granada, [221, 239, 240, 242] University of Illinois at Urbana Champaign, 189, 191, 217, 223, 235, 362, 363, 364, 383, 389] University of Nebraska Lincoln, 108, 322] Patents 13 University of Strathclyde, 390, 391, 392] University of Sussex, 156, 159, 161, 165, 365, 366, 367, 368, 369, 370, 371, 372, 373, 374, 376] ....

....370, 371, 372, 373, 374, 375, 376] Hashiyama, Tomonori, 43] Haupt, Randy L. 127] Haupt, Sue Ellen, 127] Heitk otter, J org, 20] Hekanaho, Jukka, 248] Hendriks, C. F. W. 44] Heraj arvi, Juha, 324] Hernandez, F. S. 89] Hern andez, Filiberto Santos, 61, 62, 114] Herrera, Francisco, [160, 176, 183, 193, 198, 221, 225, 232, 239, 240, 241, 242, 243, 244, 245, 246, 252, 253, 275, 276, 283, 285, 286, 288, 290, 292, 377, 378, 379] Herrera Viedma, E. 160] Herrera Viedmai, E. 239, 285] Higgins, Desmond G. 73] Hill, A. 45] H ohfeld, Markus, 254] H ohn, Christian, 255] Holland, John H. 380, 381] Honavar, Vasant, 382] Horn, Je rey, 383] H orner, Helmut, 80] Hou, Edwin S. H. 85] Hraber, Peter T. 387, 388] ....

[Article contains additional citation context not shown here]

Francisco Herrera, Manuel Lozano, and Jose Luis Verdegay. A learning process for fuzzy control rules using genetic algorithms. Technical Report DECSAI-95108, University of Granada, Department of Computer Science and Articial Intelligence, 1995. (decsai.ugr.es: pub/tech rep/ga-fl/ LP-FCR-GA.ps.Z) Key: ga95fHerrera.

.... Universitat der Berlin, 171, 308, 309, 352] The University of Texas at Austin, 349] Universidad de Granada, 142, 203, 204, 332] University of Bristol, 211] University of California at Berkley, 18] University of Cambridge, 53] University of Durham, 351] University of Granada, [181, 199, 200, 202] University of Illinois at Urbana Champaign, 148, 150, 177, 183, 195, 317, 318, 319, 338, 344] University of Nebraska Lincoln, 85, 282] University of Strathclyde, 345, 346, 347] University of Sussex, 114, 117, 119, 123, 320, 321, 322, 323, 324, 325, 326, 327, 328, 329, 331] University of ....

....Stephen J. 341] Harvey, Inman, 119, 73, 320, 321, 322, 323, 324, 325, 326, 327, 328, 329, 330, 331] Hashiyama, Tomonori, 33] Haupt, Randy L. 97] Haupt, Sue Ellen, 97] Heitkotter, Jorg, 19] Hekanaho, Jukka, 208] Hendriks, C. F. W. 34] Herajarvi, Juha, 283] Herrera, Francisco, [118, 135, 142, 152, 157, 181, 185, 192, 199, 200, 201, 202, 203, 204, 205, 206, 212, 213, 235, 236, 243, 245, 246, 248, 250, 252, 332, 333, 334] Herrera Viedma, E. 118] Herrera Viedmai, E. 199, 245] Higgins, Desmond G. 54] Hill, A. 35] Hohfeld, Markus, 214] Hohn, Christian, 215] Holland, John H. 335, 336] Honavar, Vasant, 337] Horn, Jeffrey, 338] Horner, Helmut, 61] Hou, Edwin S. H. 65] Hraber, Peter T. 342, 343] ....

[Article contains additional citation context not shown here]

Francisco Herrera, Manuel Lozano, and Jose Luis Verdegay. A learning process for fuzzy control rules using genetic algorithms. Technical Report DECSAI-95108, University of Granada, Department of Computer Science and Artificial Intelligence, 1995. (decsai.ugr.es: pub/tech rep/ga-fl/ LP-FCR-GA.ps.Z) Key: ga95fHerrera.

....have been developed successfully using Fuzzy Logic [1, 13, 15] Learning algorithms are This work has been supported by CICYT under Project TIC95 0453 useful in order to obtain automatically fuzzy controllers using examples from the system. In this field, several articles may be found [14, 20, 21]. Genetic algorithms have shown themselves to be a very useful tool for the construction of learning algorithms [3, 4, 6] In this paper, we describe SLAVE, a system of learning from examples, that uses a genetic algorithm and concepts of fuzzy technology. This system is able to cope with fuzzy ....

Herrera F., Lozano M., Verdegay J.L., A learning process for Fuzzy Control Rules using Genetic Algorithms, Technical Report #DECSAI-95108 (1995)

....chromosomes are generated with the original values of the DB in CSP and CSA parts, and alleles at random (within the set f0; 1; 2g) in the CSL part. The crossover operator will depend on the chromosome part where it is applied: In CSP and CSA parts, the max min arithmetical crossover [12] is considered. In the CSL part, the standard two point crossover is used. After recombining each part, the two best chromosomes among the eight (four di erent CSP and CSA parts combined with two di erent CSL parts) descendants obtained will be selected to replace their parents. The ....

F. Herrera, M. Lozano, J.L. Verdegay, A learning process for fuzzy control rules using genetic algorithms, Fuzzy Sets and Systems 100 (1998) 143-158.

.... Rule reduction methods have been formulated using Neural Networks, clustering techniques and orthogonal transformation methods, similarity measures [46, 10, 23, 47, 34, 53, 55, 33, 48] as well as using GA based rule selection processes to get a co operative set of rules from a candidate rule set [31, 16, 26, 24, 15, 36, 45]) From a di erent point of view, in [11] an attempt to reduce the growth of the rule set by proposing a disjunctive form for the fuzzy rules (a rule combination method) is considered. In [13] a multistage genetic learning process was presented, based on the MOGUL methodology [15] that deals ....

....Systems The multistage genetic learning process for FRBCSs consists of three stages [13] 1. An iterative fuzzy rule generation process, that obtains a linguistic RB which represents the knowledge extracted from the training examples and veri es the completeness and k consistency properties [20, 24]. 2. A genetic multiselection process, that generates di erent KBs by the selection of di erent rule subsets and the learning of di erent linguistic modi er sets, considering the FRM used in the classi cation stage. 3. A genetic tuning process, that leads to obtain the best membership function ....

[Article contains additional citation context not shown here]

F. Herrera, M. Lozano, and J. L. Verdegay. A learning process for fuzzy control rules using genetic algorithms. Fuzzy Sets and Systems, 100:143-158, 1998.

....FRB in this stage. Simplified FRBs presenting the best possible cooperation between the fuzzy rules composing them could be obtained. The method proposed in [10] will be considered for this task allowing the use of TSK rules. In the mentioned work, the genetic simplification process proposed in [17] is used as basic algorithm of unimodal optimization, and a sequential niche technique [6] is used to generate niches in the search space allowing the method to obtain different FRBs. 3. Genetic tuning stage. This process adjusts the antecedent membership function definitions of the fuzzy rules ....

....as part of a multicriteria fitness function used to evaluate the rules in the fuzzy rule generating method. All told, the following three frequentistic criteria are considered in the fitness function for an approximate fuzzy rule, R i , through the set of examples, EN : ffl High frequency value [17]: The frequency of a fuzzy rule, R i , through the set of examples, EN , is defined as: Psi EN (R i ) P N l=1 R i (e l ) N ; with R i (e l ) being the covering degree of the rule R i over the example e l . ffl High average covering degree over positive examples [17] Being the set of ....

[Article contains additional citation context not shown here]

F. Herrera, M. Lozano, J.L. Verdegay, A learning process for fuzzy control rules using genetic algorithms, Fuzzy Sets and Systems 100 (1998) 143--158.

....by the linguistic variables. The genetic multiselection process obtains different simplified KBs, with the best co operation between the rules. It includes: The Sequential Niche Technique [6] to induce niches [20] using as basic optimisation technique the genetic selection process proposed in [35], iterated in each run of the multiselection process. A search process that looks for the best set of modifiers or linguistic hedges associated with the linguistic labels of the variables. A local search posterior to each selection process, so that for the best individual, i.e. the best KB, ....

Herrera, F., Lozano M. and Verdegay, J.L. (1998), "A learning process for fuzzy control rules using genetic algorithms," Fuzzy Sets and Systems, Vol. 100, pp. 143-158.

....supported by DGICYT PB92 0933 fuzzy rule generation Wang and Mendel s method by using them to develop a fuzzy modeling of a threedimensional control surface derived from a mathematical function. The method proposed consists of the following three steps, maintaining the generic structure used in [10, 5, 6]: 1. An evolutionary generation process for generating fuzzy control rules, with two components: a fuzzy rule generating method based on an inductive algorithm with an Evolution Strategy that locally tunes the rules, and an iterative covering method of the system behaviour example set. 2. A ....

....the constrains imposed over the membership functions. When the membership function shapes are constrained by an initial domain fuzzy partition, then we may say that the rules present constrained free semantic [5, 6] On the other hand, when no restriction is imposed, an unconstrained free semantic [10, 6] is considered. In the generation process presented in this paper we work with this last approach. 4 The Genetic Generation Process As it has been commented, the first stage consists of two processes, a generating method of desirable fuzzy rules from examples and a covering method of the set of ....

[Article contains additional citation context not shown here]

Herrera, F., Lozano, M., Verdegay, J.L.: A Learning Process for Fuzzy Control Rules Using Genetic Algorithms. Technical Report DECSAI-- 95108, Dept. of Computer Science and A.I., University of Granada, Spain (February 1995).

....labels. This component allows us to obtain a set of Mamdani type fuzzy rules B g describing the system behavior. In order to do that, it is necessary to establish a condition for it. This is the requirement of covering all possible situation action pairs, e l 2 E p , the completeness property [15, 48]. This may be formalized for a constant 2 [0; 1] it requires the non zero union of fuzzy sets A i ( B i ( i = 1; T , T = jB g j, and is formulated by the following expressions: CR (e l ) i=1: T R i (e l ) l = 1; p R i : If x 1 is A i1 and . and x n is A in ....

....in the generation process, the accuracy of the candidates is measured by using a multicriteria tness function. This function is designed taking into account the following three criteria allowing us to ensure the completeness and consistency of the nal KB generated: 23 a) High frequency value [48] The frequency of a fuzzy rule, R i , through the set of examples, E p , is de ned as: Ep (R i ) P p l=1 R i (e l ) p : b) High average covering degree over positive examples [48] The set of positive examples to R i with compatibility degree greater than or equal to is de ned as: ....

[Article contains additional citation context not shown here]

Herrera, F., Lozano, M., and Verdegay, J. L. (1998). A learning process for fuzzy control rules using genetic algorithms. Fuzzy Sets and Systems 100, pp. 143-158.

....them classi ed according to the said aspects. Table 1: Analyzed Learning Methods Ref. Author(s) Method Technique Learning [4] B ardossy, Duckstein WCA AHDD HCL Ant UL Cons [22, 6] Dunn, extended by Bezdek FCM Clustering UL [9] Chiu CEB Clustering UL [8] Carse, Fogarty, Munro P FCS1 GA UL [27] Herrera, Lozano, Verdegay MOGUL GLP GA UL [15] Cord on, Herrera MOGUL SCL GA SCL [17] Cord on, Herrera MOGUL HCL GA HCL AHDD = Ad Hoc Data Driven, GA = Genetic Algorithm, HCL = Hard Constrained Learning, SCL = Soft Constrained Learning, UL = Unconstrained Learning, Ant = Antecedent, Cons = ....

....features of GAs make them suitable candidates to incorporate prior knowledge (fuzzy membership function parameters, fuzzy rules, number of rules, etc. Over the last few years, these advantages have extended the use of GAs in the development of a wide range of approaches for designing FRBSs [8, 11, 15, 16, 17, 27]. Three alternative approaches have been proposed to apply GAs to learning processes: the Michigan [31] the Pittsburgh [37] and the Iterative Rule Learning (IRL) 41] approaches. In the rst one, the chromosomes correspond to classi er rules that are evolved as a whole, whereas in the ....

[Article contains additional citation context not shown here]

F. Herrera, M. Lozano, J.L. Verdegay, A learning process for fuzzy control rules using genetic algorithms, Fuzzy Sets and Systems 100 (1998) 143-158.

.... of the individuals is performed using the stochastic universal sampling procedure together with an elitist selection scheme, and the generation of the offspring population is put into effect by using the classical binary multipoint crossover (performed at two points) and uniform mutation operators [4, 8]. The coding scheme generates fixed length chromosomes. Considering the rules contained in the linguistic rule set derived from the previous step counted from 1 to m, an m bit string C = c 1 ; c m ) represents a subset of candidate rules to form the KB finally obtained as this stage output, ....

F. Herrera, M. Lozano, J.L. Verdegay, A learning process for fuzzy control rules using genetic algorithms, Fuzzy Sets and Systems (1998, to appear).

....in terms of the fuzzy sets involved in them. This is likely to be of benefit in tackling the course of dimensionality when scaling to multi dimensional systems. Anyway, its drawback with respect to the descriptive FRBS is the loss of FRB readability. The approximate approach is considered in [4, 9, 11, 13, 21, 23, 24, 29, 33, 38, 45]. 2.2 Genetic Fuzzy Rule Based Systems EAs, especially GAs, have proven to be a powerful tool for automating the definition of the FRB, since adaptive control, learning, and self organizative FRBSs can be considered in many cases as optimization or search processes. Their advantages have ....

....Fuzzy Rule Base It is clear that an FRBS should always be able to infer a proper output for every possible system input. This property may be called completeness in the field of inductive learning and it may be mathematically formulated using a real value by means of the following expression [38]: CR (e l ) i=1: T R i (e l ) l = 1; p R i (e l ) A i1 (ex l 1 ) A in (ex l n ) B i (ey l ) where is a t norm, and R i (e l ) is the compatibility degree between the rule R i and the example e l . Given an FRB composed of T fuzzy rules R i , the covering value ....

[Article contains additional citation context not shown here]

F. Herrera, M. Lozano, J.L. Verdegay, A learning process for fuzzy control rules using genetic algorithms, Fuzzy Sets and Systems (1998, to appear).

No context found.

F. Herrera, M. Lozano, J.L. Verdegay, "A learning process for fuzzy control rules using genetic algorithms," Fuzzy Sets and Systems (1998). To appear.

....final KB as output by adjusting the mem bership functions for each fuzzy rule in each possible KB obtained from the stage above. A more complete description of the multistage GFSs may be found in [20] Different multi stage GFSs for learning KB following these ideas may be found in [10, 11, 23, 26] 6 An example of GFS This section will describe, in a few lines, one of the GFSs previously cited, specifically a GFS learning RBs using a Pittsburgh approach and representing the rule base with a decision table. This method was proposed by Philip Thrift ( 49] This example will be analyzed ....

F. Herrera, M. Lozano, J.L. Verdegay. A Learning process for fuzzy control rules using genetic algorithms. Fuzzy Sets and Systems (1997) To appear.

....to combine the advantages and solve the drawbacks presented by classical genetic learning approaches, the Michigan and Pittsburgh ones [26] when applied to the problem of designing Rule Based Systems. There are different EFSs based on the IRL in the specialized literature (see [27] 12] 28] [29]) These kind of FRBS evolutionary design processes are based on three main aspects [25] 27] 1. As in the Michigan approach, each chromosome in the population represents a single fuzzy rule, but only the best individual is considered to form part of the final KB. Therefore, in this approach the ....

....t is small, and a very local one in later stages. D.2 Crossover In this case we shall work with another genetic operator which has shown good behavior for RCGAs, the max min arithmetical crossover. This crossover operator was proposed in [32] and has been widely used in the field of EFSs [33] [29], 27] 12] It works in the way shown below. If C t v = c 1 ; c k ; c H ) and C t w = c 0 1 ; c 0 k ; c 0 H ) are to be crossed, the following four offspring are generated C t 1 1 = aC t w (1 Gamma a)C t v C t 1 2 = aC t v (1 Gamma a)C t w C t 1 3 ....

F. Herrera, M. Lozano, and J. L. Verdegay, "A learning process for fuzzy control rules using genetic algorithms," To appear in Fuzzy Sets and Systems, 1997.

....be possible to automatically generate a complete FLC KB when a training set formed by numerical input output (state control) problem variable pairs recorded experimentally is available. The method proposed consists of the following three steps, whilst maintaining the generic structure used in [8]: 1. An iterative and increasing RB generation process of desirable fuzzy control rules able to include the complete knowledge of the set of examples, 2. A genetic rule simplification process, which finds the final RB able to approximate the input output behaviour of the real system. It is based ....

....This component allows us to obtain a set of Mamdani type fuzzy control rules B g describing the system behaviour. In order to do that, it is necessary to establish a condition for it. This is the requirement of covering all possible situation action pairs, e l 2 E p , the completeness property [6, 8]. This may be formalized for a constant 2 [0; 1] it requires the non zero union of fuzzy sets A i ( Delta) B i ( Delta) i = 1; T , T = jB g j, and is formulated by the following expressions: CR (e l ) i=1: T R i (e l ) l = 1; p (1) R i : If x 1 is A i1 and . and xn ....

[Article contains additional citation context not shown here]

Herrera, F., Lozano, M., Verdegay, J.L.: A Learning Process for Fuzzy Control Rules Using Genetic Algorithms. Technical Report DECSAI--95108, Dept. of Computer Science and A.I., University of Granada, Spain (February 1995).

....a single rule, but contrary to the latter, only the best individual is considered as the solution, discarding the remaining chromosomes in the population. Therefore, in the iterative model, the GA provides a partial solution to the problem of learning. This model has been used in papers such as [42, 16, 18, 19, 24, 25, 9] and attempts to reduce the search space for the possible solutions. In order to obtain a set of rules, which will be a true solution to the problem, the GA has to be placed within an iterative scheme similar to the following: 1. Use a GA to obtain a rule for the system. 2. Incorporate the rule ....

....KB, fuzzy rules and membership functions. We find approaches presenting variable chromosome length, others coding a fixed number of rules and their membership functions, several working with chromosomes encoding single control rules instead of a complete KBs, etc. Some approaches are presented in [6, 34, 37, 25, 41, 4, 9]. For a more detailed description see [8] for an extensive bibliography see [7] section 3.13) and some approaches may be found in [27] In the following, we present the MSGFS for learning RB or KB based on the iterative rule learning approach. 4.2 A Multi Stage Genetic Fuzzy System Learning ....

[Article contains additional citation context not shown here]

Herrera, F., Lozano, M., Verdegay, J.L., A Learning Process for Fuzzy Control Rules using Genetic Algorithms. Tech. Report #DECSAI \Gamma 95108, Dept. of Computer Science and A.I., University of Granada, 1995.

....KB, removing from it the rules not cooperating well. The main idea of the genetic multisimplification process is that it does not only generate one simplified definition of the previous fuzzy rule set, but several different ones. To do so, it runs the genetic simplification process proposed in [23]. This process is based on a binary coded GA which encodes the set of rules obtained from the generation process into a fixed length chromosome. The value 1 means that the rule belongs to the final KB, and the 0 means that it does not. Two point crossover and uniform mutation operators are used to ....

F. Herrera, M. Lozano, J.L. Verdegay, "A learning process for fuzzy control rules using genetic algorithms," Fuzzy Sets and Systems, 1998, to appear.

No context found.

F. Herrera, M. Lozano, and J.L. Verdegay, "A learning process for fuzzy control rules using genetic algorithms.," Fuzzy Sets and Systems, vol. 100, pp. 143--158, 1998.

No context found.

F. Herrera, M. Lozano, and J. Verdegay, "A Learning Process for Fuzzy Control Rules using Genetic Algorithms," Department of Computer Science and A.I., University of Granada, 18071 Granada, Spain, Tech. Rep. DECSAI-95108, February 1995.

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