Sanghoun Oh

A genetic-inspired multicast routing optimization algorithm with bandwidth and end-to-end delay constraints

This paper presents a genetic-inspired multicast routing algorithm with Quality of Service (i.e., bandwidth and end-to-end delay) constraints. The aim is to efficiently discover a minimum-cost multicast tree (a set of paths) that satisfactorily helps various services from a designated source to multiple destinations. To achieve this goal, state of the art genetic-based optimization techniques are employed. Each chromosome is represented as a tree structure of Genetic Programming. A fitness function that returns a tree cost has been suggested. New variation operators (i.e., crossover and mutation) are designed in this regard. Crossover exchanges partial chromosomes (i.e., sub-trees) in a positionally independent manner. Mutation introduces (in part) a new sub-tree with low probability. Moreover, all the infeasible chromosomes are treated with a simple repair function. The synergy achieved by combing new ingredients (i.e., representation, crossover, and mutation) offers an effective search capability that results in improved quality of solution and enhanced rate of convergence. Experimental results show that the proposed GA achieves minimal spanning tree, fast convergence speed, and high reliability. Further, its performance is better than that of a comparative reference.

Incremental approximation of nonlinear constraint functions for evolutionary constrained optimization

This paper proposes an alternative approach to efficient solving of nonlinear constrained optimization problems using evolutionary algorithms. It is assumed that the separate-ness of the feasible regions, which imposes big difficulties for evolutionary search, is partially resulted from the complexity of the nonlinear constraint functions. Based on this hypothesis, an approximate model is built for each constraint function with an increasing accuracy, starting from a simple linear approximation. As a result, the feasible region based on the approximate constraint functions will be much simpler, and the isolated feasible regions will become more likely connected. As the evolutionary search goes on, the approximated feasible regions should gradually change back to the original one by increasing the accuracy of the approximate models to ensure that the optimum found by the evolutionary algorithm does not violate any of the original constraints. Empirical studies have been performed on 13 test problems and four engineering design optimization problems. Simulation results suggest that the proposed method is competitive compared to the state-of-the-art techniques for solving nonlinear constrained optimization problems.

An evolutionary cluster validation index

This paper presents a new evolutionary method for the cluster validation index (CVI), namely eCVI. The proposed method learns CVI from the generated training data set using the genetic programming (GP), and then outputs the optimal number of clusters after taking parameters of a test data set into the learned CVI. Each chromosome encodes a possible CVI as a function of the number of clusters, density measure of clusters, and some random factors. Fitness function evaluating each candidate is defined by the difference between the actual number of clusters from training data set and the number of clusters computed by the current CVI. Because of the adaptive nature of GP, the proposed eCVI is reliable and robust in various types of data sets. Experimental results provide grounds for the dominance of eCVI over several widely-known CVIs.

Evolutionary optimization programming with probabilistic models

Genetic programming is a powerful optimization technique thanks to its capacity of discovering automatically a proper set of programs, rules or functions of a given problem. Regardless of such strengths, GP does not handle a key genetic operator, crossover effectively, resulting in the disruption of good building blocks. To overcome such a problem, we propose a probabilistic model-based evolutionary optimization programming in this paper. It utilizes an enhanced expanded parse tree that transforms the tree into linear-type chromosomes by inserting nulls and selectors, and that reduces the size of a conditional probability table. Also, a multivariate dependence model, χ-ary extended compact genetic algorithm, χ-eCGA, is employed to find a good probability distribution in the form of marginal product model for the problem. Experimental results provide grounds for the dominance of the proposed approach over existing algorithms.

On the practical genetic algorithms

This paper offers practical design-guidelines for developing efficient genetic algorithms (GAs) to successfully solve real-world problems. As an important design component, a practical population-sizing model is presented and verified.

Incremental Approximation Models for Constrained Evolutionary Optimization

Many real-world scientific and engineering problems are constrained optimization problems (COPs). To solve these problems, a variety of evolutionary algorithms have been proposed by incorporating different constraint-handling techniques. However, many of them have difficulties in achieving the global optimum due to the presence of highly constrained feasible regions in the search space. To effectively address the low degree of feasibility, this chapter presents an incremental approximation strategy-assisted constraint-handling method in combination with a multi-membered evolution strategy. In the proposed approach, we generate an approximate model for each constrained function with increasing accuracy, from a linear-type approximation to a model that has a complexity similar to the original constraint functions, thereby manipulating the complexity of the feasible region. Thanks to this property, our constrained evolutionary optimization algorithm can acquire the optimal solution conceivably. Simulations are carried out to compare the proposed algorithm with well-known references on 13 benchmark problems and three engineering optimization problems. Our computational results demonstrate that the proposed algorithm is comparable or superior to the state of the art on most of the test problems used in this study and a spring design optimization problem.

Optimal Total Exchange in Anonymous Cayley Graphs

Total exchange or all-to-all personalized communication problem is that each node in a network has a message to be sent to every other node. To solve this communication problem, we present a time-optimal algorithm in anonymous Cayley graphs as assuming a single-port full duplex model in which every node is able to send and receive at most one message in each time unit.