Andreas L. Symeonidis

Intelligent policy recommendations on enterprise resource planning by the use of agent technology and data mining techniques

Abstract Enterprise Resource Planning systems tend to deploy Supply Chain Management and/or Customer Relationship Management techniques, in order to successfully fuse information to customers, suppliers, manufacturers and warehouses, and therefore minimize system-wide costs while satisfying service level requirements. Although efficient, these systems are neither versatile nor adaptive, since newly discovered customer trends cannot be easily integrated with existing knowledge. Advancing on the way the above mentioned techniques apply on ERP systems, we have developed a multi-agent system that introduces adaptive intelligence as a powerful add-on for ERP software customization. The system can be thought of as a recommendation engine, which takes advantage of knowledge gained through the use of data mining techniques, and incorporates it into the resulting company selling policy. The intelligent agents of the system can be periodically retrained as new information is added to the ERP. In this paper, we present the architecture and development details of the system, and demonstrate its application on a real test case.

A Framework for Constructing Multi-agent Applications and Training Intelligent Agents

As agent-oriented paradigm is reaching a significant level of acceptance by software developers, there is a lack of integrated high-level abstraction tools for the design and development of agent-based applications, In an effort to mitigate this deficiency, we introduce Agent Academy, an integrated development framework, implemented itself as a multi-agent system, that supports, in a single tool, the design of agent behaviours and reusable agent types, the definition of ontologies, and the instantiation of single agents or multi-agent communities. In addition to these characteristics, our framework goes deeper into agents, by implementing a mechanism for embedding rule-based reasoning into them. We call this procedure agent training and it is realized by the application of AI techniques for knowledge discovery on application-specific data, which may be available to the agent developer. In this respect, Agent Academy provides an easy-to-use facility that encourages the substitution of existing, traditionally developed applications by new ones, which follow the agent-orientation paradigm.

A robust agent design for dynamic SCM environments

The leap from decision support to autonomous systems has often raised a number of issues, namely system safety, soundness and security. Depending on the field of application, these issues can either be easily overcome or even hinder progress. In the case of Supply Chain Management (SCM), where system performance implies loss or profit, these issues are of high importance. SCM environments are often dynamic markets providing incomplete information, therefore demanding intelligent solutions which can adhere to environment rules, perceive variations, and act in order to achieve maximum revenue. Advancing on the way such autonomous solutions deal with the SCM process, we have built a robust, highly-adaptable and easily-configurable mechanism for efficiently dealing with all SCM facets, from material procurement and inventory management to goods production and shipment. Our agent has been crash-tested in one of the most challenging SCM environments, the trading agent competition SCM game and has proven capable of providing advanced SCM solutions on behalf of its owner. This paper introduces Mertacor and its main architectural primitives, provides an overview of the TAC SCM environment, and discusses Mertacors performance.

Agent Mertacor: A robust design for dealing with uncertainty and variation in SCM environments

Supply Chain Management (SCM) has recently entered a new era, where the old-fashioned static, long-term relationships between involved actors are being replaced by new, dynamic negotiating schemas, established over virtual organizations and trading marketplaces. SCM environments now operate under strict policies that all interested parties (suppliers, manufacturers, customers) have to abide by, in order to participate. And, though such dynamic markets provide greater profit potential, they also conceal greater risks, since competition is tougher and request and demand may vary significantly in the quest for maximum benefit. The need for efficient SCM actors is thus implied, actors that may handle the deluge of (either complete or incomplete) information generated, perceive variations and exploit the full potential of the environments they inhabit. In this context, we introduce Mertacor, an agent that employs robust mechanisms for dealing with all SCM facets and for trading within dynamic and competitive SCM environments. Its efficiency has been extensively tested in one of the most challenging SCM environments, the Trading Agent Competition (TAC) SCM game. This paper provides an extensive analysis of Mertacor and its main architectural primitives, provides an overview of the TAC SCM environment, and thoroughly discusses Mertacors performance.

Data mining for agent reasoning: A synergy for training intelligent agents

The task-oriented nature of data mining (DM) has already been dealt successfully with the employment of intelligent agent systems that distribute tasks, collaborate and synchronize in order to reach their ultimate goal, the extraction of knowledge. A number of sophisticated multi-agent systems (MAS) that perform DM have been developed, proving that agent technology can indeed be used in order to solve DM problems. Looking into the opposite direction though, knowledge extracted through DM has not yet been exploited on MASs. The inductive nature of DM imposes logic limitations and hinders the application of the extracted knowledge on such kind of deductive systems. This problem can be overcome, however, when certain conditions are satisfied a priori. In this paper, we present an approach that takes the relevant limitations and considerations into account and provides a gateway on the way DM techniques can be employed in order to augment agent intelligence. This work demonstrates how the extracted knowledge can be used for the formulation initially, and the improvement, in the long run, of agent reasoning.

BioCrawler: An intelligent crawler for the semantic web

Web crawling has become an important aspect of web search, as the WWW keeps getting bigger and search engines strive to index the most important and up to date content. Many experimental approaches exist, but few actually try to model the current behaviour of search engines, which is to crawl and refresh the sites they deem as important, much more frequently than others. BioCrawler mirrors this behaviour on the semantic web, by applying the learning strategies adopted in previous work on ecosystem simulation, called BioTope. BioCrawler employs the principles of BioTopes intelligent agents on the semantic web, learns which sites are rich in semantic content and which sites link to them and adjusts its crawling habits accordingly. In the end, it learns to behave much like the state of the art search engine crawlers do. However, BioCrawler reaches that behavior solely by exploiting on-page factors, rather than off-page factors, such as the currently used link popularity.

Event identification in web social media through named entity recognition and topic modeling

The problem of identifying important online or real life events from large textual document streams that are freely available on the World Wide Web is increasingly gaining popularity, given the flourishing of the social web. An event triggers discussion and comments on the WWW, especially in the blogosphere and in microblogging services. Consequently, one should be able to identify the involved entities, topics, time, and location of events through the analysis of information publicly available on the web, create semantically rich representations of events, and then use this information to provide interesting results, or summarize news to users. In this paper, we define the concept of important event and propose an efficient methodology for performing event detection from large time-stamped web document streams. The methodology successfully integrates named entity recognition, dynamic topic map discovery, topic clustering, and peak detection techniques. In addition, we propose an efficient algorithm for detecting all important events from a document stream. We perform extensive evaluation of the proposed methodology and algorithm on a dataset of 7million blogposts, as well as through an international social event detection challenge. The results provide evidence that our approach: a) accurately detects important events, b) creates semantically rich representations of the detected events, c) can be adequately parameterized to correspond to different social perceptions of the event concept, and d) is suitable for online event detection on very large datasets. The expected complexity of the online facet of the proposed algorithm is linear with respect to the number of documents in the data stream.

Information agents cooperating with heterogenous data sources for customer-order management

As multi-agent systems and information agents obtain an increasing acceptance by application developers, existing legacy Enterprise Resource Planning (ERP) systems still provide the main source of data used in customer, supplier and inventory resource management. In this paper we present a multi-agent system, comprised of information agents, which cooperates with a legacy ERP in order to carry out orders posted by customers in an enterprise environment. Our system is enriched by the capability of producing recommendations to the interested customer through agent cooperation. At first, we address the problem of information workload in an enterprise environment and explore the opportunity of a plausible solution. Secondly we present the architecture of our system and the types of agents involved in it. Finally, we show how it manipulates retrieved information for efficient and facile customer-order management and illustrate results derived from real-data.

Syncing Shared Multimedia through Audiovisual Bimodal Segmentation

This work emanates from the particularities residing in contemporary social media storytelling, where multiple users and publishing channels capture and share public events, experiences, and places. Multichannel presentation and visualization mechanisms are pursued along with novel audiovisual mixing (such as time-delay-compensation enhancement, perceptual mixing, quality-based content selection, linking to context-aware metadata, and propagating multimedia semantics), thus promoting multimodal social media editing, processing, and authoring. While the exploitation of multiple time-based media (audio and video) describing the same event may lead to significant content enhancement, difficulties regarding detection and temporal synchronization of multimedia events have to be overcome. In many cases, one can identify events based only on audio features, thus performing an initial cost-effective annotation of the multimedia content. This article introduces a new audio-driven approach for temporal alignment and management of shared audiovisual streams. The article presents the theoretical framework and demonstrates the methodology in real-world scenarios. This article is part of a special issue on social multimedia and storytelling.

A retraining methodology for enhancing agent intelligence

Data mining has proven a successful gateway for discovering useful knowledge and for enhancing business intelligence in a range of application fields. Incorporating this knowledge into already deployed applications, though, is highly impractical, since it requires reconfigurable software architectures, as well as human expert consulting. In an attempt to overcome this deficiency, we have developed agent academy, an integrated development framework that supports both design and control of multiagent systems (MAS), as well as agent training. We define agent training as the automated incorporation of logic structures generated through data mining into the agents of the system. The increased flexibility and cooperation primitives of MAS, augmented with the training and retraining capabilities of agent academy, provide a powerful means for the dynamic exploitation of data mining extracted knowledge. In this paper, we present the methodology and tools for agent retraining. Through experimental results with the agent academy platform, we demonstrate how the extracted knowledge can be formulated and how retraining can lead to the improvement - in the long run - of agent intelligence.