Gordan Ježić

Semantics in Multi-Agent Systems

In this chapter we discuss how semantic technologies in general and specific Semantic Web standards in particular can contribute to the goal of achieving interoperability between independent, loosely coupled, heterogeneous, autonomous software components (i.e. agents) and for the realization of open interaction systems. In particular we will discuss how those technologies have been used for the definition of the semantics of agent communication languages, for the definition of norms and policies used to regulate interactions in open frameworks, and for defining efficient mechanisms for matching demands (i.e., content they need) to supplies (i.e., available content) in telecommunication networks. In particular regarding this last type of application we describe a techno-economic approach for solving the matching problem, by means of a multi-agent system representing an electronic marketplace. Its functionality is realized by applying a semantic-aware content discovery model with two-level filtering in order to finally recommend a ranked set of eligible content to the users in response to their requests. The filtering processes not only consider the semantic information associated with the available content, but also ratings regarding the actual performance of businesses that act as content providers as well as the prices paid by businesses for advertising their content.

A high throughput processing engine for taxi-generated data streams

The ACM DEBS Grand Challenge 2015 focuses on real-time analytics over a high volume geospatial data stream composed of taxi trip reports from New York City. The goal of the challenge is to provide a solution which continuously identifies the most frequent routes (query 1) and most profitable areas (query 2) for taxis in New York City. The solution needs to process the incoming data stream in near real-time to provide valid information about taxi positions to end-users in a real-world deployment. We propose a modular processing engine design which is configured to offer efficient performance with a high data throughput and low processing latency. It consists of three main components: an input processor which pre-processes data objects to detect outliers, and two independent query processors tailored to the requirements of challenge queries. To efficiently compute query results, query processors use algorithms customized to the distribution of the taxi-generated data stream. Our experimental evaluation shows that the system can on average process 350,000 input events per second in a distributed mode, while achieving an average latency of less than 1 ms for both queries. Due to their excellent performance, the proposed algorithms are well suited for efficient tracking of a large number of vehicles that are present in modern urban areas.

Implementation of Agent-Based Games Recommendation System on Mobile Platforms

Because of Google Play and App Store, today numerous different games are offered to every smartphone user. This diversity in supply is undoubtedly a good thing, but it also virtually disables users to find games they would like to play. However, it was shown that users tend to spend more money on purchasing recommended games when these recommendations are done using personal recommenders. In this paper we present an agent-based recommender for mobile platforms in which recommendations are made taking into account user game experience. Inputs for our recommender, which are collected both inside and outside the games, are stored in a semantic database. Based on collected information, user and game profiles are made that are then used in our recommendation algorithm. The focus of the paper is on how to do the implementation of the proposed system in real-world environments and obtain all the necessary data and how to make recommendations based on generated user and game profiles.