Experiments toward a practical implementation of an intelligent kanban system

This paper presents laboratory experiments to test a bottom up approach to production control and supply chain management. Built upon the successful traditional kanban (Card) system, the new intelligent system associates a kanban agent to each physical kanban. Instead of relying on demand forecast and planning, kanban agents reason about their own movements to adapt to changing demands. After previous simulations results of the intelligent system showed significant performance improvements over the traditional system, we further use the Auto-ID Laboratory at Cambridge University to test the feasibility of the idea in a realistic manufacturing environment. The results from the experiments demonstrated the superiority on several performance measures of the intelligent system compared to the traditional system used as a benchmark. Moreover, the implementation of the experiments exposed several real world constraints not shown in the simulation study and practical solutions were adopted to address these.

Enhancements to fraud and deception prevention in open multi-agent marketplaces

Fraud and deception in online marketplaces has been an on-going problem. This thesis proposes novel techniques and mechanisms using agent technology to protect buyers and sellers in online environments such as eBay. The proposed solution has been rigorously tested and the results show good commercial promise.

Agent-based heuristics to solve NP-hard scheduling problems

A novel agent-driven heuristic approach was developed to control the operational scheduling for a local manufacturer. This approach outperformed the traditional kanban control mechanism under numerous simulated benchmarking tests. Using this approach, the individual machine loading was reduced by, on average, 28%, with the loading spread reduced by 85%

Co-ordination in multi-agent systems

A multi-agent system is a complex software system which is composed of many relative autonomous smaller softwares called agents. The research on multi-agent systems is concerned with the interaction and coordination among these agents to let them help each other to solve complicated problems, such as finance investment management. The principal contributions represented by these 50 selected papers are "cooperation under uncertainty in distributed expert systems (DESs)", "a tool and algorithms to build DESs", and "information gathering and decision making in multi-agent systems (MASs)".

CARSA : A context-aware reasoning-based service agent model for AI planning of web service compostion

In order to achieve automatic and more intelligent service composition, dynamic description logic (DDL) is proposed and utilized as one emerging logic-level solution. However, reasoning optimization and utilization in such DDL-related solutions is still an open problem. In this paper, we propose the context-aware reasoning-based service agent model (CARSA) which exploits the relationships among different service consumers and providers, together with the corresponding optimization approach to strengthen the effectiveness of Web service composition. Through the model, two reasoning optimization methods are proposed based on the substitute relationship and the dependency relationship, respectively, so irrelevant actions can be filtered out of the reasoning space before the DDL reasoning process is carried out. The case study and experimental analysis demonstrates the capability of the proposed approach.

Intelligent traffic light control of isolated intersections using machine learning methods

Traffic congestion is one of the major problems in modern cities. This study applies machine learning methods to determine green times in order to minimize in an isolated intersection. Q-learning and neural networks are applied here to set signal light times and minimize total delays. It is assumed that an intersection behaves in a similar fashion to an intelligent agent learning how to set green times in each cycle based on traffic information. Here, a comparison between Q-learning and neural network is presented. In Q-learning, considering continuous green time requires a large state space, making the learning process practically impossible. In contrast to Q-learning methods, the neural network model can easily set the appropriate green time to fit the traffic demand. The performance of the proposed neural network is compared with two traditional alternatives for controlling traffic lights. Simulation results indicate that the application of the proposed method greatly reduces the total delay in the network compared to the alternative methods.

Ethical trust and social moral norms simulation : a bio-inspired agent-based modelling approach

The understanding of the micro-macro link is an urgent need in the study of social systems. The complex adaptive nature of social systems adds to the challenges of understanding social interactions and system feedback and presents substantial scope and potential for extending the frontiers of computer-based research tools such as simulations and agent-based technologies. In this project, we seek to understand key research questions concerning the interplay of ethical trust at the individual level and the development of collective social moral norms as representative sample of the bigger micro-macro link of social systems. We outline our computational model of ethical trust (CMET) informed by research findings from trust, machine ethics and neural science. Guided by the CMET architecture, we discuss key implementation ideas for the simulations of ethical trust and social moral norms.

Distributed multi-agent scheme for reactive power management with renewable energy

This paper presents a new distributed multi-agent scheme for reactive power management in smart coordinated distribution networks with renewable energy sources (RESs) to enhance the dynamic voltage stability, which is mainly based on controlling distributed static synchronous compensators (DSTATCOMs). The proposed control scheme is incorporated in a multi-agent framework where the intelligent agents simultaneously coordinate with each other and represent various physical models to provide information and energy flow among different physical processes. The reactive power is estimated from the topology of distribution networks and with this information, necessary control actions are performed through the proposed proportional integral (PI) controller. The performance of the proposed scheme is evaluated on a 8-bus distribution network under various operating conditions. The performance of the proposed scheme is validated through simulation results and these results are compared to that of conventional PI-based DSTATCOM control scheme. From simulation results, it is found that the distributed MAS provides excellence performance for improving voltage profiles by managing reactive power in a smarter way.

Agent-based smart grid protection and security

The vision of a smart grid is to provide a modern, resilient, and secure electric power grid as it boasts up with a highly reliable and efficient environment through effective use of its information and communication technology (ICT). Generally, the control and operation of a smart grid which integrate the distributed energy resources (DERs) such as, wind power, solar power, energy storage, etc., largely depends on a complex network of computers, softwares, and communication infrastructure superimposed on its physical grid architecture facilitated with the deployment of intelligent decision support system applications. In recent years, multi-agent system (MAS) has been well investigated for wide area power system applications and specially gained a significant attention in smart grid protection and security due to its distributed characteristics. In this chapter, a MAS framework for smart grid protection relay coordination is proposed, which consists of a number of intelligent autonomous agents each of which are embedded with the protection relays. Each agent has its own thread of control that provides it with a capability to operate the circuit breakers (CBs) using the critical clearing time (CCT) information as well as communicate with each other through high speed communication network. Besides physical failure, since smart grid highly depends on communication infrastructure, it is vulnerable to several cyber threats on its information and communication channel. An attacker who has knowledge about a certain smart grid communication framework can easily compromise its appliances and components by corrupting the information which may destabilize a system results a widespread blackout. To mitigate such risk of cyber attacks, a few innovative counter measuring techniques are discussed in this chapter.

A decentralised multi-agent approach to enhance the stability of smart microgrids with renewable energy

This paper presents the impact of large penetration of wind power on the transient stability through a dynamic evaluation of the critical clearing times (CCTs) by using intelligent agent-based approach. A decentralised multi-agent-based framework is developed, where agents represent a number of physical device models to form a complex infrastructure for computation and communication. They enable the dynamic flow of information and energy for the interaction between the physical processes and their activities. These agents dynamically adapt online measurements and use the CCT information for relay coordination to improve the transient stability of power systems. Simulations are carried out on a smart microgrid system for faults at increasing wind power penetration levels and the improvement in transient stability using the proposed agent-based framework is demonstrated.

Reducing the gap between mobile and intelligent software agents

This paper presents an approach to integrate an artificial intelligence (AI) technique, concretely rule-based processing, into mobile agents. In particular, it focuses on the aspects of designing and implementing an appropriate inference engine of small size to reduce migration costs. The main goal is combine two lines of agent research, First, the engineering oriented approach on mobile agent architectures, and, second, the AI related approach on inference engines driven by rules expressed in a restricted subset of first-order predicate logic (FOPL). In addition to size reduction, the main functions of this type of engine were isolated, generalized and implemented as dynamic components, making possible not only their migration with the agent, but also their dynamic migration and loading on demand. A set of classes for representing and exchanging knowledge between rule-based systems was also proposed.

Saccharomyces cerevisiae: a novel and efficient biological control agent for Colletotrichum acutatum during pre-harvest

In this study, we evaluated the efficiency of six isolates of Saccharomyces cerevisiae in controlling Colletotrichum acutatum, the causal agent of postbloom fruit drop that occur in pre-harvest citrus. We analyzed the mechanisms of action involved in biological control such as: production of antifungal compounds, nutrient competition, detection of killer activity, and production of hydrolytic enzymes of the isolates of S. cerevisiae on C. acutatum and their efficiency in controlling postbloom fruit drop on detached citrus flowers. Our results showed that all six S. cerevisiae isolates produced antifungal compounds, competed for nutrients, inhibited pathogen germination, and produced killer activity and hydrolytic enzymes when in contact with the fungus wall. The isolates were able to control the disease when detached flowers were artificially inoculated, both preventively and curatively. In this work we identified a novel potential biological control agent for C acutatum during pre-harvest. This is the first report of yeast efficiency for the biocontrol of postbloom fruit drop, which represents an important contribution to the field of biocontrol of diseases affecting citrus populations worldwide. (C) 2015 Elsevier GmbH. All rights reserved.

Engineering Agent-Oriented Technologies and Programming Languages for Computer Programming and Software Development

Mainstream hardware is becoming parallel, heterogeneous, and distributed on every desk, every home and in every pocket. As a consequence, in the last years software is having an epochal turn toward concurrency, distribution, interaction which is pushed by the evolution of hardware architectures and the growing of network availability. This calls for introducing further abstraction layers on top of those provided by classical mainstream programming paradigms, to tackle more effectively the new complexities that developers have to face in everyday programming. A convergence it is recognizable in the mainstream toward the adoption of the actor paradigm as a mean to unite object-oriented programming and concurrency. Nevertheless, we argue that the actor paradigm can only be considered a good starting point to provide a more comprehensive response to such a fundamental and radical change in software development. Accordingly, the main objective of this thesis is to propose Agent-Oriented Programming (AOP) as a high-level general purpose programming paradigm, natural evolution of actors and objects, introducing a further level of human-inspired concepts for programming software systems, meant to simplify the design and programming of concurrent, distributed, reactive/interactive programs. To this end, in the dissertation first we construct the required background by studying the state-of-the-art of both actor-oriented and agent-oriented programming, and then we focus on the engineering of integrated programming technologies for developing agent-based systems in their classical application domains: artificial intelligence and distributed artificial intelligence. Then, we shift the perspective moving from the development of intelligent software systems, toward general purpose software development. Using the expertise maturated during the phase of background construction, we introduce a general-purpose programming language named simpAL, which founds its roots on general principles and practices of software development, and at the same time provides an agent-oriented level of abstraction for the engineering of general purpose software systems.

Knowledge-Based Task Structure Planning for an Information Gathering Agent

An effective solution to model and apply planning domain knowledge for deliberation and action in probabilistic, agent-oriented control is presented. Specifically, the addition of a task structure planning component and supporting components to an agent-oriented architecture and agent implementation is described. For agent control in risky or uncertain environments, an approach and method of goal reduction to task plan sets and schedules of action is presented. Additionally, some issues related to component-wise, situation-dependent control of a task planning agent that schedules its tasks separately from planning them are motivated and discussed.

Multi-agent Architecture for Heterogeneous Reasoning under Uncertainty Combining MSBN and Ontologies in Distributed Network Diagnosis

This article proposes a MAS architecture for network diagnosis under uncertainty. Network diagnosis is divided into two inference processes: hypothesis generation and hypothesis confirmation. The first process is distributed among several agents based on a MSBN, while the second one is carried out by agents using semantic reasoning. A diagnosis ontology has been defined in order to combine both inference processes. To drive the deliberation process, dynamic data about the influence of observations are taken during diagnosis process. In order to achieve quick and reliable diagnoses, this influence is used to choose the best action to perform. This approach has been evaluated in a P2P video streaming scenario. Computational and time improvements are highlight as conclusions.