Autonomous Intelligent Agents Research Articles

Controlling the learning process of real-time heuristic search

Real-time search provides an attractive framework for intelligent autonomous agents, as it allows us to model an agent's ability to improve its performance through experience. However, the behavior of real-time search agents is far from rational during the learning (convergence) process, in that they fail to balance the efforts to achieve a short-term goal (i.e., to safely arrive at a goal state in the present problem solving trial) and a long-term goal (to find better solutions through repeated trials). As a remedy, we introduce two techniques for controlling the amount of exploration, both overall and per trial. The weighted real-time search reduces the overall amount of exploration and accelerates convergence. It sacrifices admissibility but provides a nontrivial bound on the converged solution cost. The real-time search with upper bounds insures solution quality in each trial when the state space is undirected. These techniques result in a convergence process more stable compared with that of the Learning Real-Time A ∗ algorithm.

Anchoring in a grounded layered architecture with integrated reasoning

The GLAIR grounded layered architecture with integrated reasoning for cognitive robots and intelligent autonomous agents has been used in a series of projects in which Cassie, the SNePS cognitive agent, has been incorporated into hardware- or software-simulated cognitive robots. In this paper, we present an informal, but coherent, overview of the GLAIR approach to anchoring the abstract symbolic terms that denote an agent’s mental entities in the lower-level structures used by the embodied agent to operate in the real (or simulated) world. We discuss anchoring in the domains of: perceivable entities and properties, actions, time, and language.

Argumentation through a distributed self-stabilizing approach

Argumentation is receiving an increasing attention as a technique for practical and uncertain reasoning underlying the realization of intelligent autonomous agents. Since a decentralized organization has been proposed by several authors as an appropriate paradigm for the design of agent architectures, we propose in this article, a distributed approach to argumentation, in which several independent asynchronous processes carry out argumentation activity, by exploiting local information only. The final result of this process is the computation of the defeat status of the arguments: we devise a general distributed algorithm, which does not rely on any specific notion of defeat between arguments. The issue of coordination has been explicitly tackled by ensuring the property of self-stabilization for the algorithm. A proof of its correctness, as well as an analysis of its complexity, is provided.

Intelligent learning and control of autonomous robotic agents operating in unstructured environments

Abstract The control of autonomous intelligent robotic agent operating in unstructured changing environments includes many objective difficulties. One major difficulty concerns the characteristics of the environment that the agent should operate in. In unstructured and changing environments the inconsistency of the terrain, the irregularity of the product and the open nature of the working environment result in complex problems of identification, sensing and control. Problems can range from the effects of varying environmental conditions on the robot sensors and traction performance through to the need to deal with the presence of unexpected situations. Another major challenge is the large amounts of uncertainty that characterises real-world environments. On the one hand, it is not possible to have exact and complete prior knowledge of these environments. On the other hand, knowledge acquired through sensing is affected by uncertainty and imprecision. The quality of sensor information is influenced by sensor noise, the limited field of view, the conditions of observation, and the inherent difficulty of the perceptual interpretation process. Because environments and users of systems continuously change, robotic agents have to be adaptive. Intelligence helps because it gives systems the capacity to adapt more rapidly to environmental changes or to handle much more complex functions. In his paper we introduce this special issue and introduce the difficulty robots are facing in unstructured environments and how learning and computational intelligence can help the robots to adapt and give them the necessary intelligence they to face the challenges they encounter in their environments.

Optimal routing of multiple-load AGV subject to LIFO loading constraints

When simple automated guided vehicles (AGVs) having no random access load transfer mechanism are used for carrying multiple loads between workstations, the loads cannot be handled independently. This paper considers the case when loads are placed in flat pallets and each new picked up pallet is loaded on the top of batch of pallets already carried by the AGV. To avoid use of excessive space and time needed to reorder pallets in the batch, the loading–unloading procedures should be performed in accordance with last-in-first-out (LIFO) rule. In this paper we formulate the condition of existence of AGV routes in which, it visits each workstation only once and meets LIFO constraint. We also suggest an algorithm for finding the shortest one among such routes. Examples are provided to illustrate the performance of the algorithm. Scope and purpose The loads transfer task is defined by an oriented graph G=( V, E) in which vertices v i (1⩽i⩽N) correspond to loading/unloading nodes in a network. Arc ( v i , v j )∈ E corresponds to existence of load to be transferred from node v i to node v j . The distance between each pair of nodes is defined by a matrix D={d(v i,v j), 1⩽i⩽N, 1⩽j⩽N} . An autonomous agent has to deliver loads in accordance with graph G. The loads are stored in the stack of the agent and are loaded and unloaded using last-in-first-out (LIFO) rule. To fulfill the transfer task the agent has to pass through all the existing vertices in some order that ensures that for each arc ( v i , v j )∈ E vertex v i is visited before vertex v j . Conditions of existence of a sequence of vertices which allow all the loads to be transferred without violating LIFO rule and visiting each vertex only once are formulated. An algorithm for finding the shortest tour meeting LIFO constraints is developed. The possible applications of the algorithm are scheduling automated guided multiple load vehicles or routing autonomous intelligent agents with stack memory organization in computer networks.

URGING DESIRE: MOTIVATIONAL MECHANISMS FOR INTELLIGENT AGENTS WITH MINDS OF THEIR OWN

In this article, we examine ways to organize an intelligent autonomous agent's goal system to simulate the behavior of a human worker in an enterprise modelling system called goal orientated multi-agent simulation environment (GOMASE). In GOMASE, an agent is supplied with a goal hierarchy, activity digraph, and a motivational mechanism in an attempt to simulate the goal's activation and attempted satisfaction. The attractiveness of this approach is that all the goals possessed by the agent are related via the ultimate life goals (i.e., self-preservation and species preservation). By using GOMASE to model enterprises, the enterprise modeller can gain an insight into the complex behavior of human workers and how they set, prioritize, and execute their goals. In this article, we also provide several sample runs of our system, with a brief explanation on the agent's behaviors.

Intelligent Autonomous Agents for Manufacturing

The paper presents a hybrid control architecture that combines the distributed hierarchical control and the autonomous agents based control concepts. The aim of this approach is to obtain a control system capable of an optimal functioning in normal conditions and robust at perturbations. The optimal control policies are off-line synthesized by a higher hierarchical layer (called supervisor) based on a discrete event model of the process and on the DES supervision theory. The robustness is ensured by the lower hierarchical level, functioning in an agent based mode

A framework for distributed agent-based engineering design support

Concurrent engineering draws together team working and cooperation, with the aim of reducing the need for costly design modifications in the later stages of design and product development. However, the complexities arising in the process of design, in general, defy formal analysis and computational support is required. In providing intelligent computational support for concurrent engineering, a combination of various problem-solving strategies may be required for complex design situations. It is proposed that an appropriate model for such provision is in the form of a set of interacting autonomous intelligent agents, possessing different problem-solving capabilities and differing degrees of intelligence. The requirements for such agent-based design support are discussed and possibilities for its realisation, with the aid of Web technology, in a distributed design environment are explored.

Using intelligent agents for pro-active, real-time urban intersection control

The usability and effectiveness of traffic control systems greatly depends on its ability of reacting upon traffic patterns and permutations. In this research we investigate the applicability of autonomous intelligent agents in Urban Traffic Control (UTC), and why these artificial intelligent strategies are useful in UTC. We propose a system that autonomously can adapt itself, based upon internal rules and its environment, at changing environments. Both long term changes as well as short term changes are accounted for in different manner. The UTC model is primarily based on several Intelligent Traffic Signalling Agents (ITSA) and some authority agents. This approach enables us to use a UTC system, based on agent technology, capable of responding to traffic conditions in real-time. The system can maintain its integrity and stability within the overall transportation system. Due to its pro-active behaviour we are able to make better use of the capacity of intersections.

COMMAS (COndition Monitoring Multi-Agent System)

The application of intelligent systems for data interpretation and condition monitoring is an advancing field of research. In recent years autonomous intelligent agents and multi-agent systems have gained much attention within different real time applications. The novel idea of COMMAS (COndition Monitoring Multi-Agent System) introduces a hierarchical decentralised multi-agent architecture developed for data interpretation and condition monitoring applications. By definition condition monitoring is concerned with detecting and distinguishing faults occurring in plant that is being monitored [1]; therefore the early diagnosis and identification of faults has a number of benefits (improvement in the plant economy, reduction in operational costs, improving the level of safety etc). A variety of intelligent techniques have been applied in plant monitoring, which resulted in the development of centralised approaches for condition monitoring, e.g., Knowledge Based Systems (KBS) [2], Model Based Reasoning (MBR) Systems [3], Case Based Reasoning (CBR) Systems [4], Artificial Neural Networks (ANN) [5] etc. These approaches tend to be fixed, so they lack flexibility and extensibility. Moving to an agent-based architecture allows simultaneous complex tasks to be performed in real-time; better handling of inaccurate data is achieved and each agent can be independently updated.

Evolutionary adaptation of dispatching agents in heterarchical manufacturing systems

We propose a new approach to job flow adaptive operational control in advanced manufacturing systems. The major feature of the method is the distribution of the control tasks among completely autonomous intelligent agents. Namely, agents are implicitly coordinated by a nature-analogous adaptation mechanism, which continuously tunes the free parameters of the control law of each agent. The proposed approach is effective and reactive to severe disturbances and changes in the manufacturing environment. Simulation experiments illustrate the operational distributed approach and its response to faults.

Intelligent Travel Planning: A MultiAgent Planning System to Solve Web Problems in the e-Tourism Domain

This paper presents Intelligent Travel Planning (ITP), a multiagent planning system to solve Web electronic problems in the Web, whose main goal is to search for useful solutions in the electronic-Tourism domain to system users. The system uses different types of intelligent autonomous agents whose main characteristics are cooperation, negotiation, learning, planning and knowledge sharing. Obviously the information used by the intelligent agents is heterogeneous and geographically distributed, since the main information source of the system is Internet. Other information sources are agent knowledge bases in the distributed system. The process to obtain, filter, and store the information is performed automatically by agents. This information is translated into a homogeneous format for high-level reasoning in order to obtain different partial solutions. Partial solutions are reconstructed into a general solution (or solutions) to be presented to the user. The system will show a set of solutions to the users that can be evaluated by them.

Soft computing based multi-agent marketing decision support system

The new concept of soft computing based distributed multi-agent marketing decision support system (DSS) is proposed in this paper. The DSS is composed of a number of parallel intelligent agents acting in a distributed mode, and one coordinating agent for evaluation of solutions. Generation of a final solution in DSS is achieved through cooperative and competitive interactions among intelligent agents. The main idea of the proposed DSS is based on decomposition of the overall system intelligence among cooperative autonomous intelligent agents capable of competing and cooperating with each other in order to propose total solution to the problem and to integrate local solutions into the final solution. Each distributed intelligent agent is implemented as fuzzy knowledge based system with modest number of rules. The agent-coordinator (evaluator) evaluates the proposals and ranks the fuzzy values of the outcome produced by intelligent contending agents. The outcome of each agent is estimated on the basis of its solution, input information, and other necessary information using a neuro-fuzzy estimator tuned by a genetic algorithm. The proposed concept for designing DSS is illustrated through a marketing-mix DSS.

Cross-paradigm analysis of autonomous agent architecture

Autonomous agent architectures are design methodologies—collections of knowledge and strategies which are applied to the problem of creating situated intelligence. This article attempts to integrate this knowledge across several architectural traditions, paying particular attention to features which have tended to be selected under the pressure of extensive use in real-world systems. We determine that the following strategies provide significant assistance in the design of autonomous intelligent agents: (i) modularity, which simplifies both design and control; (ii) hierarchically organized action selection, which focusses attention and provides prioritization when different modules conflict; and (iii) parallel environment monitoring which allows a system to be responsive and opportunistic by allowing attention to shift and priorities to be re-evaluated. We offer a review of four architectural paradigms: behaviour-based AI; two- and three-layered systems; belief, desire and intention architectures (particularly PRS); and Soar/ACT-R. By documenting trends within each of these communities towards establishing the components above, we argue that this convergent evolution is strong evidence for the components' utility. We then use this information to recommend specific strategies for researchers working under each paradigm to further exploit the knowledge and experience of the field as a whole.

Guided Missile with an Intelligent Agent

Guided missiles involve the use of a conventional deviated pursuit course like proportional navigation algorithm and its variants, which is optimal when the speed advantage of the guided missile is very high and the target maneouvering is minimal. Against the present-day aircraft, which employs fly-by-wire technology for high maneouverability and high speed, missiles need to have a much higher speed advantage or to use a combination of artificial intelligence and modern control algorithms. Results of Simulation of pursuit and evasion with an autonomous intelligent agent incorporated in the control loop are presented.

Multi-agent distributed intelligent system based on fuzzy decision making

Generally decision making for solving ill-structured problems in DSS takes place in uncertain situations. The main drawbacks of existing traditional DSS are inefficiencies associated with dealing with complex models and large databases. Usually a fuzzy DSS has many input variables and, hence, its knowledge base, containing the totality of fuzzy rules, is very large. Large rule base leads to disadvantages in speed, reliability, and complexity of DSS. This paper introduces an alternative concept for designing fuzzy DSS based on multi-agent distributed artificial intelligent technology and fuzzy decision making. The main idea of the proposed DSS is based on granulation of the overall system intelligence between cooperative autonomous intelligent agents capable of competing and cooperating with each other in order to propose a total solution to the problem and organization (combining individual solutions) of the proposed solution into the final solution. It is supposed that every agent in DSS is characterized by a set of fuzzy criteria of unequal importance and definition of a “winner” agent is based on multi-criteria fuzzy decision making involving unequal objectives. © 2000 John Wiley & Sons, Inc.

Multi-Agent Systems: A Survey from the Robot-Soccer Perspective

ABSTRACTOne of the most challenging goals in artificial intelligence (AI) is the development of artificial intelligent autonomous agents with human-level performance. The past few years has witnessed a tremendous interest in research and discussions on intelligent agents. With the ever increasing number of robots in an industrial environment, scientists and technologists are often faced with issues on cooperation and coordination among different robots and their self governance in a common work-space. This has led to developments in multi-robot cooperative autonomous systems. With an aim to study issues such as group architecture, resource conflict, origin of cooperation, learning and geometric problems, groups of robots are constructed. The proponents of multi-robot autonomous systems needed a model to validate the theories being proposed and to test their efficacy and efficiency. It is no surprise that they started focussing on robot soccer. Robot soccer makes heavy demands in all the key areas of robot...

Self-evolutionary intelligent manufacturing systems based on machine learning

The present paper describes in detail about self-learning and self-evolution in Intelligent Manufacturing Systems. Along with materials being converted to finished products, systems synthesize information into knowledge by self-learning. After a kind of parts passes through systems, self-evolution begins, by means of acquired knowledge, functionality of subsystems and units in system is improved, and organization of systems is reconfigurated in order to reach the global optimization. Intelligent autonomous agents are applied to implement two mechanisms, this can make systems steadily become more and more smart and clever during manufacturing of parts. Finally, a study case about Micro-CIMS are investigated.

Behavior‐Based Robotics

Behavior‐Based Robotics

Active mental entities

In order to design and realize intelligent autonomous agents, a very active research trend has concentrated its attention on the study of models of mental activity, encompassing the explicit representation of mental attitudes such as beliefs, desires and intentions. In such studies mental attitudes are normally represented as data structures on which a so-called interpreter operates determining the overall agent behavior. In this article we propose an original point of view about mental activity modeling, founded on two basic claims: i) mental attitudes should be regarded as autonomous active entities; ii) an intelligent agent should be conceived as distributed structure, where global behavior is produced by interactions among active mental entities.