Abstracts


Swarm Intelligence (PSO + ACO)

Swarm intelligence is concerned with the design of intelligent multiagent systems by taking inspiration from the collective behavior of social insects such as ants, termites, bees, and wasps, as well as from other animal societies such as flocks of birds or fish schools. Colonies of social insects have mesmerized researchers for many decades, and the mechanisms that govern their behavior remained unknown for a long time. Even though the single members of these colonies are non-sophisticated individuals, they are able to achieve complex tasks in cooperation. Coordinated colony behavior emerges from relatively simple actions or interactions between the colonies' individual members.
Optimization techniques based on swarm intelligence have become increasingly popular during the last decade. They are characterized by a decentralized way of working that mimics the behavior of swarms of social insects, flocks of birds, or fish schools. The advantage of these approaches over traditional techniques is their robustness and flexibility. These properties make swarm intelligence a successful design paradigm for algorithms that deal with increasingly complex problems. After a general introduction, we will focus in this tutorial on two of the most successful examples of optimization techniques inspired by swarm intelligence: ant colony optimization, inspired by the foraging behavior of real ant colonies, and particle swarm optimization, inspired by bird flocking.

Transportation and Logistics

Transportation plays an important role in many industries as well as in daily life. The increase of global trade and growing energy costs lead to an increasing demand for efficient transportation and logistics. We present an overview of problems arising in transportation and review applications of evolutionary computation and other metaheuristics in the context of logistics. Like in other problem domains, the success of evolutionary algorithms depends on the choice of representation and search space, locality of operators and integration of local search. Problems studied in academia typically differ significantly from real-world problems. We give an overview of a commercial transportation management system and its planning capabilities. The system allows solving real-world vehicle routing problems and employs evolutionary local search as solution approach. We contrast the real-world vehicle routing problem from its academic counterparts and give an outlook for challenging problems of high practical relevance, which represent promising research opportunities in the area of transportation, logistics and metaheuristics.

Computational Complexity of Evolutionary Computation in Combinatorial Optimization

Randomized search heuristics such as evolutionary algorithms or ant colony optimization have been shown to be very successful when dealing with real-world applications or problems from combinatorial optimization. In recent years, a lot of progress has been made in understanding the runtime behavior of these heuristics from a theoretical point of view.
This is an important research area where a lot of interesting questions are still open. In this tutorial, we would like to give an overview how to analyze randomized search heuristics with respect to the number of solutions they need to produce until a good one has been found. We start by considering the optimization of artificial pseudo-boolean functions and point out some general methods. After that, we consider some well-known combinatorial optimization problems such as minimum spanning trees, Eulerian cycles, and the NP-hard partition problem and show how the runtime behavior of different randomized search heuristics can be analyzed for these problems.

A Unified Approach to EC

The field of Evolutionary Computation has experienced tremendous growth over the past 20 years, resulting in a wide variety of evolutionary algorithms and applications. The result poses an interesting dilemma for many practitioners in the sense that, with such a wide variety of algorithms and approaches, it is often hard to se the relationships between them, assess strengths and weaknesses, and make good choices for new application areas.
This tutorial is intended to give an overview of a general EC framework that can help compare and contrast approaches, encourages crossbreeding, and facilitates intelligent design choices. The use of this framework is then illustrated by showing how traditional EAs can be compared and contrasted with it, and how new EAs can be effectively designed using it.
Finally, the framework is used to identify some important open issues that need further research.

Bioinformatics

Bioinformatics is concerned with trying to make sense of the huge and growing amount of biological data that is emerging from modern biotechnology. We have immense amounts of sequence information about genes and proteins, lots of partial information about the roles and functions of proteins and many other biomolecules, and lots of information that connects these data with disease and other state information about biological systems. But, to be able to extract useful knowledge and tools from these data, optimisation, machine learning, and statistics expertise is essential. Even more essential, however, is for there to be scientists who understand both the biology and the advanced computer science. The tutorial will assume that you know about optimisation and machine learning, but you want to learn what the opportunities are for applying your knowledge in the bioinformatics arena. So, the tutorial will start with some relevant biology basics, and then move on to discuss and explore some of the classification and optimisation problems that need your expertise.

Evolution Strategies and Related Estimation of Distribution Algorithms (ES feat. EDA)

Evolution Strategies and some continuous domain Estimation of Distribution Algorithms are stochastic optimization procedures based on sampling multivariate Gaussian (normal) distributions. They can be formulated in a common, unified, but still very simple framework. Such a framework is very useful to understand subtle differences of algorithms.
This tutorial focuses on the most relevant question: how should the parameters of the sample distribution be chosen, and how should they be updated in the generation sequence? The most common and successful approaches are reviewed, including self-adaptation, success rules, path length control, Covariance Matrix Adaptation (CMA), and Estimation of Multivariate Normal Algorithm (EMNA).
Methods are motivated with respect to the difficulties one has to face when solving continuous domain non-linear, non-convex optimization problems. Specific problem difficulties will be discussed in the beginning, for example ill-conditioning and non-separability. Finally, the performance of state-of-the art Evolution Strategies is related to other well-known evolutionary and non-evolutionary optimization algorithms, namely BFGS, DE, PSO,...

Multi-Objective Evolutionary Algorithms (EMOA)

Multiple, often conflicting objectives arise naturally in most real-world optimization scenarios. As evolutionary algorithms possess several characteristics that are desirable for this type of problem, this class of search strategies has been used for multiobjective optimization for more than a decade. Meanwhile evolutionary multiobjective optimization has become established as a separate subdiscipline combining the fields of evolutionary computation and classical multiple criteria decision making.
This tutorial gives an overview of evolutionary multiobjective optimization with the focus on methods and theory. On the one hand, basic principles of multiobjective optimization are presented, and various algorithmic aspects such as fitness assignment and environmental selection are discussed in the light of state-of-the-art techniques. On the other hand, the tutorial covers several theoretical issues such as performance assessment and running-time analysis.

Games

This tutorial provides a practical introduction to game strategy learning with function approximation architectures. The tutorial will cover the two main approaches to learning game strategy: evolution (including co-evolution), and temporal difference learning.
We also look at how the choice of input features and function approximation architecture has a critical impact on what is learned, as well as how it is interfaced to the game (e.g. as a value estimator or as an action selector). In addition to standard MLPs, attention is also given to N-Tuple systems and interpolated table-based approximators, as these have recently shown great potential to learn quickly and effectively.
Each method will be demonstrated with reference to some simple fragments of software, illustrating how the learning algorithm is connected with the game and with the function approximation architecture. Example games will include Othello, Simulated Car Racing, and Ms. Pac-Man.
Recent results on the information rate limits attainable for these learning algorithms (in terms of bits of information gained per game played) will also be discussed, together with a simple test game that enables direct testing of these information rates.

The Future of Experimental Research

It is an open secret that the performance of algorithms depends on their parameterizations --- and of the parameterizations of the problem instances. However, these dependencies can be seen as means for understanding algorithm's behavior. Based on modern statistical techniques we demonstrate how to tune and understand algorithms. We present a comprehensive, effective and very efficient methodology for the design and experimental analysis of search heuristics such as evolutionary algorithms, differential evolution, pattern search or even classical deterministic methods such as the Nelder-Mead simplex algorithm. Our approach extends the sequential parameter optimization (SPO) method that has been successfully applied as a tuning procedure to numerous heuristics for practical and theoretical optimization problems. Optimization practitioners receive valuable hints for choosing an adequate heuristic for their optimization problems --- theoreticians receive guidelines for testing results systematically on real problem instances. Based on several examples from theory and practice we demonstrate how SPO improves the performance of many search heuristics significantly. However, this performance gain is not available for free. Therefore, costs of this tuning process are discussed.