CEC

This paper describes a new approach towards optimum selection of the different parameters of the mould oscillation system in the continuous casting process of steel manufacturing. The objective of optimization is to enhance lubrication within the mould, especially at higher casting speeds, and reduce the intensity of oscillation marks. The need to improve lubrication conditions is primarily addressed by making the mould oscillate on its longitudinal axis. It is known that non-sinusoidal oscillation, where the time for upward motion of the mould is longer than that of downward motion in an oscillation cycle, reduces depth of oscillation marks while providing better lubrication. In the present work, a Genetic Algorithm is applied to optimize the amplitude, frequency, and waveform of the oscillation of the continuous casting mould based on objective functions that maximize the lubrication, and minimize the depth of oscillation marks and the cycle peak friction. Optimization is performed within constraints imposed by machine limits. The objective function and constraints are extracted from an analysis of the physics of oscillation, lubrication and heat transfer within the continuous casting process. The application of the Genetic Algorithm within a unified framework encompassing all oscillation performance metrics and constraints is seen to generate an optimal parameter set that provides better performance than existing oscillation parameters supplied by Original Equipment Manufacturers.

This paper presents Natural Evolution Strategies (NES), a recent family of algorithms that constitute a more principled approach to black-box optimization than established evolutionary algorithms. NES maintains a parameterized distribution on the set of solution candidates, and the natural gradient is used to update the distribution's parameters in the direction of higher expected fitness. We introduce a collection of techniques that address issues of convergence, robustness, sample complexity, computational complexity and sensitivity to hyperparameters. This paper explores a number of implementations of the NES family, ranging from general-purpose multi-variate normal distributions to heavy-tailed and separable distributions tailored towards global optimization and search in high dimensional spaces, respectively. Experimental results show best published performance on various standard benchmarks, as well as competitive performance on others.

In this paper we employ a steady state genetic algorithm to evolve different types of behaviour for bots in the Unreal Tournament 2004™ computer game. For this purpose we define three fitness functions which are based on the number of enemies killed, the lifespan of the bot and a combination of both. Long run experiments were carried out, in which the evolved bots' behaviours outperform those of standard bots supplied by the game, particularly in those cases where the fitness involves a measure of the bot's lifespan. Also, there is an increase in the number of items collected, and the behaviours tend to become less aggressive, tending instead towards a more optimised combat style. Further “short run” experiments were carried out with a further type of fitness function defined, based on the number of items picked. In these cases the bots evolve performances towards the goal they have been aimed, with no other behaviours arising, except in the case of the multiple objective one. We conclude that in order to evolve interesting behaviours more complex fitness functions are needed, and not necessarily ones that directly include the goal we are aiming for.

Efficient vehicle path planning in hostile environment to carry out rescue or tactical logistic missions remains very challenging. Most approaches reported so far relies on key assumptions and heuristic procedures to reduce problem complexity. In this paper, a new model and a hybrid genetic algorithm are proposed to solve the rescue path planning problem for a single vehicle navigating in uncertain adversarial environment. We present a simplified mathematical linear programming formulation aimed at minimizing traveled distance and threat exposure. As an approximation to the basic problem, the user-defined model allows to specify a lower bound on the optimal solution for some particular survivability conditions. Hard problem instances are then solved using a novel hybrid genetic algorithm relaxing some of the common assumptions considered by previous path construction methods. The algorithm evolves a population of solution combining genetic operators with a new stochastic path generation technique, providing guided local search, while improving solution quality. The value of the problem-solving approach is shown for simple cases and compared to an alternate heuristic.

Popular games often have a high-quality graphic design but quite simple-minded non player characters (NPC). Recently, Computational Intelligence (CI) methods have been discovered as suitable methods to revive NPC, making games more interesting, challenging, and funny. We present a fairly large study of human players on the simple arcade game Pac-Man, controlling the ghosts behaviors by simple strategies, neural networks or evolutionary algorithms. The playerpsilas fun is of course a subjective experience, but we presume that it is related to the psychological flow concept. We deal with the question whether flow is a more reliable measure than asking human players directly for the fun experienced during the game. In order to detect flow, we introduce a measure based on the interaction time fraction between the human-controlled Pac-Man and the ghosts, and compare the outcome to the results of a fun measure suggested by Yannakakis and Hallam [1].