International Journal of Computational Intelligence and Applications

Any sufficiently expressive construction kit — expert systems, decision trees, neural networks, to name a few — with enough design work, can perform practically any set of behaviors. And when the required behaviors are not known, interviewing human experts (if available) is a workable but labor-intensive method.

In this work a co-evolutionary approach is used in conjunction with Genetic Programming operators in order to find certain transition rules for two-step discrete dynamical systems. This issue is similar to the well-known artificial-ant problem. We seek the dynamic system to produce a trajectory leading from given initial values to a maximum of a given spatial functional.

This problem is recast into the framework of input-output relations for controllers, and the optimization is performed on program trees describing input filters and finite state machines incorporated by these controllers simultaneously. In the context of Genetic Programming there is always a set of test cases which has to be maintained for the evaluation of program trees. These test cases are subject to evolution here, too, so we employ a so-called host-parasitoid model in order to evolve optimizing dynamical systems.

Reinterpreting these systems as algorithms for finding the maximum of a functional under constraints, we have derived a paradigm for the automatic generation of adapted optimization algorithms via optimal control. We provide numerical examples generated by the GP-system MathEvEco. These examples refer to key properties of the resulting strategies and they include statistical evidence showing that for this problem of system identification the co-evolutionary approach is superior to standard Genetic Programming.

Perhaps the best known example of user-guided evolution is furnished by evolving expressions, an image generation technique first introduced by Sims. In this version of artificial evolution, images are evolved for aesthetic purposes, hence any fitness measure used must be based on aesthetics. We consider the problem of guiding image evolution autonomously on the basis of computational, as opposed to user-assigned, aesthetic fitness. Due to the difficulty of formulating an absolute criterion for aesthetic fitness, we adopt a coevolutionary approach, relying on hosts and parasites to establish relative criteria for aesthetic fitness. To sustain the coevolutionary arms race, we allow coevolution to proceed in stages. This permits appropriate fitness levels to be maintained within the parasite populations we use to infect host image populations. Using staged coevolution produces two beneficial results: (1) longer survival times for subpopulations of host images, and (2) stable phenotypic lineages for host images.

In this paper, we will propose a novel framework of hybridization of Coevolutionary Genetic Algorithm and Machine Learning. The Coevolutionary Genetic Algorithm (CGA) which has already been proposed by Handa et al. consists of two GA populations: the first GA (H-GA) population searches for the solutions in given problems, and the second GA (P-GA) population searches for effective schemata of the H-GA. The CGA adopts the notion of commensalism, a kind of co-evolution. The new hybrid framework incorporates a schema extraction mechanism by Machine Learning techniques into the CGA. Considerable improvement in its search ability is obtained by extracting more efficient and useful schemata from the H-GA population and then by incorporating those extracted schemata into the P-GA. We will examine and compare two kinds of machine learning techniques in extracting schema information: C4.5 and CN2. Several computational simulations on multidimensional knapsack problems, constraint satisfaction problems and function optimization problems will reveal the effectiveness of the proposed methods.

Previous work on coevolutionary search has demonstrated both successful and unsuccessful applications. As a step in explaining what factors lead to success or failure, we present a comparative study of an evolutionary and a coevolutionary search model. In the latter model, strategies for solving a problem coevolve with training cases. We find that the coevolutionary model has a relatively large efficacy: 86 out of 100 (86%) of the simulations produce high quality strategies. In contrast, the evolutionary model has a very low efficacy: a high quality strategy is found in only two out of 100 runs (2%). We show that the increased efficacy in the coevolutionary model results from the direct exploitation of low quality strategies by the population of training cases. We also present evidence that the generality of the high-quality strategies can suffer as a result of this same exploitation.

An interesting problem which has been widely investigated is under what circumstances will a society of rational agents realize some particular stable situations, and whether they satisfy the condition of social efficiency? This will crucially depend on how they interact and what information they have when they interact. For instance, when strategic interactions are modeled as coordination games, it is known the evolutionary process selects the risk-dominant equilibrium which is not necessarily efficient. We consider the networks of agents, in which each agent faces several types of the strategic decision problems. We investigate the dynamics of collective decision when each agent adapts the strategy of interaction to its neighbors. We are interested in to show how the society gropes its way towards an equilibrium situation. We show that the society selects the most efficient equilibrium among multiple equilibria when the agents composing it do learn from each other as collective learning, and they co-evolve their strategies over time. We also investigate the mechanism that leads the society to an equilibrium of social efficiency.

There is a widespread perception that in conflict situations, more intermediate choices between full peace and total war makes full peace less likely. This view is a motivation for opposing the proposed National Missile Defense. This perception is partly due to research in the abstract game of Iterated Prisoner's Dilemma. This paper critically evaluates this perception.