Narrow Results

In this paper, a novel genetic algorithm — dynamic ring-like agent genetic algorithm (RAGA) is proposed for solving global numerical optimization problem. The RAGA combines the ring-like agent structure and dynamic neighboring genetic operators together to get better optimization capability. An agent in ring-like agent structure represents a candidate solution to the optimization problem. Any agent interacts with neighboring agents to evolve. With dynamic neighboring genetic operators, they compete and cooperate with their neighbors, and they can also use knowledge to increase energies. Global numerical optimization problems are the most important ones to verify the performance of evolutionary algorithm, especially of genetic algorithm and are mostly of interest to the corresponding researchers. In the corresponding experiments, several complex benchmark functions were used for optimization, several popular GAs were used for comparison. In order to better compare two agents GAs (MAGA: multi-agent genetic algorithm and RAGA), the several dimensional experiments (from low dimension to high dimension) were done. These experimental results show that RAGA not only is suitable for optimization problems, but also has more precise and more stable optimization results.

As Bayesian networks become widely accepted as a normative formalism for diagnosis based on probabilistic knowledge, they are applied to increasingly larger problem domains. These large projects demand a systematic approach to handle the complexity in knowledge engineering. The needs include modularity in representation, distribution in computation, as well as coherence in inference. Multiply Sectioned Bayesian Networks (MSBNs) provide a distributed multiagent framework to address these needs.

According to the framework, a large system is partitioned into subsystems and represented as a set of related Bayesian subnets. To ensure exact inference, the partition of a large system into subsystems and the representation of subsystems must follow a set of technical constraints. How to satisfy these goals for a given system may not be obvious to a practitioner. In this paper, we address three practical modeling issues.

Autonomy is an often cited but rarely agreed upon agent characteristic. Although no definition of agent autonomy is universally accepted, the concept of adaptive autonomy promises increasingly flexible and robust agent-based systems. In general, adaptive autonomy gives agents the ability to seek help for problems or take initiative when otherwise they would be constrained by their design to follow some fixed procedures or rules for interacting with other agents. In order to access these benefits, this article provides a core definition and representation of agent autonomy designed to support the implementation of adaptive agent autonomy. This definition identifies "decision-making control" governing the determination of agent goals and tasks as the key dimension of agent autonomy. In order to gain run-time flexibility and any associated performance improvements, agents must be able to dynamically adapt their autonomy during system operation. This article justifies the implementation of dynamic adaptive autonomy through a series of experiments showing that a multiagent system operating under dynamic adaptive autonomy performs better than a multiagent system operating under fixed autonomy for the same changing run-time conditions.

This article proposes a novel multiagent approach to optimization inspired by diffusion in nature called Evolutionary Multiagent Diffusion (EMD). Each agent in EMD makes the decision to diffuse based on the information shared between its parent and its siblings. The behavior of EMD is analyzed and its relation to similar search algorithms is discussed.

In this paper, we extend and modify the ERA approach proposed in Ref. 13 to solve Propositional Satisfiability Problems (SATs). The new ERA approach involves a multiagent system where each agent only senses its local environment and applies some self-organizing rules for governing its movements. The environment, which is a two-dimensional cellular environment, records and updates the local values that are computed and affected according to the movements of individual agents. In solving a SAT with the ERA approach, we first divide variables into several groups, and represent each variable group with an agent whose possible positions correspond to the elements in a Cartesian product of variable domains, and then randomly place each agent onto one of its possible positions. Thereafter, the ERA system will keep on dispatching agents to choose their movements until an exact or approximate solution emerges. The experimental results on some benchmark SAT test-sets have shown that the ERA approach can obtain comparable results as well as stable performances for SAT problems. In particular, it can find approximate solutions for SAT problems in only a few steps. The real value of this approach is that it is a distributed asynchronous approach without any centralized control or evaluation, where the agents can cooperate to solve problems without explicit communication.

Multiagent systems (MAS) provide a useful tool for exploring the complex dynamics and behavior of financial markets and now MAS approach has been widely implemented and documented in the empirical literature. This paper introduces the implementation of an innovative multi-scale mathematical model for a computational agent-based financial market. The paper develops a method to quantify the degree of self-organization which emerges in the system and shows that the capacity of self-organization is maximized when the agent behaviors are heterogeneous. Numerical results are presented and analyzed, showing how the global market behavior emerges from specific individual behavior interactions.

This paper introduces the implementation of a computational agent-based financial market model in which the system is described on both microscopic and macroscopic levels. This artificial financial market model is used to study the system response when a shock occurs. Indeed, when a market experiences perturbations, financial systems behavior can exhibit two different properties: resilience and robustness. Through simulations and different scenarios of market shocks, these system properties are studied. The results notably show that the emergence of collective herding behavior when market shock occurs leads to a temporary disruption of the system self-organization. Numerical simulations highlight that the market can absorb strong mono-shocks but can also be led to rupture by low but repeated perturbations.

Emergence of game strategy in multiagent systems is studied. Symbolic and subsymbolic (neural network) approaches are compared. Symbolic approach is represented by a backtrack algorithm with specified search depth, whereas the subsymbolic approach is represented by feedforward neural networks that are adapted by reinforcement temporal difference TD(λ) technique. As a test game, we use simplified checkers. The problem is studied in the framework of multiagent system, where each agent is endowed with a neural network used for a classification of checkers positions. Three different strategies are used. The first strategy corresponds to a single agent that repeatedly plays games against MinMax version of a backtrack search method. The second strategy corresponds to single agents that are repeatedly playing a megatournament, where each agent plays two different games with all other agents, one game with white pieces and the other game with black pieces. After finishing each game, both agents modify their neural networks by reinforcement learning. The third strategy is an evolutionary modification of the second one. When a megatournament is finished, each agent is evaluated by a fitness, which reflects its success in the given megatournament (more successful agents have greater fitness). It is demonstrated that all these approaches lead to a population of agents very successfully playing checkers against a backtrack algorithm with the search depth 3.

This paper is concerned with near-optimal source search problem using a multiagent system in cluttered indoor environments. The goal of the problem is to maximize the detection probability within the minimum search time. We propose a two-stage strategy to achieve this goal. In the first stage, a greedy approach is used to define a set of grid cells with the aim of maximizing the detection probability. In the second stage, an iterative branch-and-bound procedure is used to design the search paths of all agents so that all grid cells are visited by one agent and the largest search path among all agents is minimized. Simulation results show that the proposed search algorithm has better performance in terms of exploration time compared to other existing methods.

The case-based reasoning (CBR) approach consists of retrieving solutions from similar past problems and adapting them to new problems. Interpolation tools can easily be used as adaptation tools in CBR systems. The accuracies of interpolated results depend on the set of known solved problems with which the interpolation tools are previously trained. EquiVox is a CBR-based system designed for retrieving and adapting three-dimensional numerical representations of human organs called phantoms. EquiVox uses an interpolation tool as an adaptation process. These phantoms are used by radiation protection experts to establish dosimetric reports in case of accidental overexposure to radiation. In addition, medical physicians need these phantoms to compute and control exposure to radiation used to treat diseases such as cancerous tumors in hospitals. The present work aims at proposing a distributed architecture for EquiVox so that a user may find a solution as quickly as possible based on the most recent available knowledge of a given community. We have designed a distributed architecture based on a multiagent paradigm and studied the theoretical performance of the new version. The ability of the new system to quickly provide and adapt solutions using the most up-to-date knowledge has been analyzed from a probabilistic angle. In the case of limited and accidental exposure to radiation, the proposed parallel processing system improves the previous and sequential version of EquiVox. Improvements are also obtained in some cases of massive exposure to radiation.

DPOP is an efficient algorithm based on the Depth First Search (DFS) Pseudo-tree for distributed constraint optimization problems in multiagent systems (MAS). DFS Pseudotree is able to help achieve parallelism due to the relative independence of nodes lying in different branches. However, we often get a chain-like pseudo-tree with few branches in our experiments, which greatly impairs the algorithm performance. Therefore we propose a new DPOP algorithm called BFSDPOP which uses Breadth First Search (BFS) Pseudotree as the communication structure. The two advantages are that BFS Pseudo-tree can help the algorithm achieve more parallelism as it has more branches; and BFS Pseudotree shortens the communication path and requires less communication time because the height of a BFS Pseudo-tree is often much lower than that of a DFS Pseudo-tree from the same constraint graph. To overcome the cross edge constraints in BFS Pseudo-tree which can easily result in large utility message size, a method of Cluster Removing is proposed. In the experiment, we compare BFSDPOP with the original DPOP and the result shows that BFSDPOP outperforms the original DPOP in most cases, which demonstrate the excellent attributes that BFS Pseudo-tree has over DFS Pseudo-tree.