Highlights of Practical Applications of Scalable Multi-Agent Systems. The PAAMS Collection

We discuss the motivation, design, implementation, and pilot study of an agent-based exercise robot for five senior residents and five staff members in a senior living community. Our goals of the study were to evaluate the performance of the resident group and the attitudes, acceptance, and opinions of both groups. The pilot study was performed on-site where senior participants were selected a priori by the staff. We use soft systems methodology as a guide to refine the requirements and to evolve the exercise protocol and robot behaviors over multiple exercise sessions. Based on a 30-min session with both groups combined, followed by focus sessions with each group individually, our findings suggest that senior residents moderately accept the robot as a group exercise leader and staff members are cautiously enthusiastic about the idea.

This paper presents our approach for addressing “Human-friendly Knowledge-INtensive Dynamic Systems” (FRIENDLY & KIND systems) from a methodological point of view, also providing tools and languages for their design, implementation and testing. FRIENDLY & KIND systems are an evolution of multiagent systems and represent a good option for engineering complex and dynamic applications like those in the e-Health domain. We will demonstrate the suitability of our approach by designing and implementing a Remote Monitoring System for oncological patients.

The clinical study of the most basic vital signs of a patient represents the simplest and most effective way to detect and monitor health problems. There are many diseases that can be diagnosed and controlled through regular monitoring of these medical data. The purpose of this study is to develop a monitoring and tracking system for the various vital signs of a patient. In particular, this work focuses on the design of a multi-agent architecture composed of virtual organizations with capabilities to integrate different medical sensors on an open, low-cost hardware platform. This system integrates hardware and software elements needed for the routine measurement of vital signs, performed by the patient or caregiver without having to go to a medical center.

One way to reduce the computational load of the agents is the distribution of their services. To achieve this goal, the functionality of a MAS (multiagent system) should not reside in the agents themselves, but ubiquitously be distributed so that allows the system to perform tasks in parallel avoiding an additional computational cost. The distribution of services that offers SCODA (Distributed and Specialized Agent Communities) allows an intelligent management of these services provided by agents of the system and the parallel execution of threads that allow to respond to requests asynchronously, which implies an improvement in the performance of the system at both the computational level as the level of quality of service in the control of these services. The comparison carried out in the case of study that is presented in this paper demonstrates the existing improvement in the distribution of services on systems based on SCODA.

Supply chain problems cover several aspects at different levels and areas. There are decision on production allocation, resource allocation, production and inventory quantities, distributor selection, choice of transportation mode etc. There are many constraints in the supply chain problems. They concern the following areas (production, distribution, transport, etc.) and types (linear, non-linear, integer, logical, etc.). Therefore it is important effective modeling and solving constraints.We consider a multi-level and multi-agent approach to modeling and solving supply chain problems using constraint and mathematical programming environments. Its efficiency results from the multi-level presolving and multi-agent architecture. An illustrative example presents effectiveness of the proposed approach. The presented approach will be compared with classical mathematical programming on the same data sets.

Modern logistics platform paradigm positively effects development of enterprises. Over time several types of logistics platforms have developed. One of them is supply chain logistics platform, which is essential to reduce the overall cost of logistic activities between supply chain partners, to improve the overall efficiency of logistics and to integrate social resources. Superficially said, it is a general information exchange platform, which applies computer, internet and communication system, and other modern information technology [1]. With the purpose of integrating social resources, it reduces the threshold of the logistics informatization and provides all kinds of accurate, timely, shared information for each participant in logistics activities [2]. Logistics platforms represent a modern approach aimed towards fostering and facilitating logistics activities and business exchange with associated flows in a specific geographic area. We examined whether stakeholders of logistics system in various types of organizations in Slovenia and Poland understand information communication system and joint interactive portal as parts of supply chain’s logistics platform. With a case study, we demonstrated that implementation of common information solution is a step on a way to contemporary and comprehensive logistics platform. Additionally, common information solution should not be simply equated with the concept of supply chain’s logistics platform. Given the important role of IT solutions in the field of supply chains, this paper proposes a novel operation procedure for coordination of supply chain actions for validation and further research.

This paper presents the use of a multi-agent framework for evaluating parameters of new products and estimating cost of product design. Companies often develop many new product projects simultaneously. A limited budget of research and development imposes selection of the most promising projects. The evaluation of new product projects requires cost estimation and involves many agents that analyse the customer requirements and information acquired from an enterprise system, including the fields of sales and marketing, research and development, and manufacturing. The model of estimating product design cost is formulated in terms of a constraint satisfaction problem. The illustrative example presents the use of a fuzzy neural network to identify the relationships and estimate cost of product design.

The paper presents a description of modelling the supply chain including disturbances by using simulation software. In order to make the best representation of reality, the route, the lorry’s speed and various types of disturbances are taken into account. The purpose of this article is to demonstrate how disturbances can be modeled and to present benefits of using the simulation programs to plan a route and time of transport.

This paper introduces an initial approach to represent human-driven, industrial-truck material handlers as agents in discrete-event simulation models of manufacturing systems. The approach is network based and involves material handlers creating work tasks for themselves based on the current states of the system, such as inventory in a production area and material availability in a supply area. The material handler integrates other work tasks with supporting the production lines. The approach leverages constructs currently available in simulation software and is implemented in FlexSim. An illustrative example is provided and the agent-based results are compared to traditional means for modeling material handling.

This paper presents the results of the project, which goal is to analyze the production process capability after reengineering the assembly process due to expansion of a bus production plant. The verification of the designed work organization for the new configuration of workstations on new production hall is necessary. The simulation model is the best tool for visualization and verification of the work organization based on individual workteams which are supporting particular workstations. Owing to the simulation it is possible to define the imperfections of this conception and elaborate improvements which will minimize the idleness of workers and downtime occurring in the assembly process. The objective of performed activities is to provide assurance that the new organization of assembly process will lead to maximum utilization of production capacity in the company. To solve described problems authors propose a method based on mixing DES (Discrete Event Simulation) and ABS (Agent Based Simulation) approach. DES was used to model the main process – material flow (buses), ABS was used to model assembling operations of teams of workers.

General-Purpose Computing on Graphics Units (GPGPU) is today recognized as a practical and efficient way of accelerating software procedures that require a lot of computing resources. However, using this technology in the context of Multi-Agent Based Simulation (MABS) appears to be difficult because GPGPU relies on a very specific programming approach for which MABS models are not naturally adapted. This paper discusses practical results from several works we have done on adapting and developing different MABS models using GPU programming. Especially, studying how GPGPU could be used in the scope of MABS, our main motivation is not only to speed up MABS but also to provide the MABS community with a general approach to GPU programming, which could be used on a wide variety of agent-based models. So, this paper first summarizes all the use cases that we have considered so far and then focuses on identifying which parts of the development process could be generalized.

We propose a holonic multiagent simulator that can simulate any complex urban environment. We focus on traffic simulation within any geographic area on earth, subject to any weather conditions. We adopt an agent-based approach for the different beahviors of the vehicles, drivers, and pedestrians. The proposed driving behavioral models can realistically emulate driving behaviors of humans. The resulting simulator can handle all the complexities of such environments in accordance with the laws of physics.

There has been an increasing interest in information diffusion within social networks and the usage of multiagent systems for knowledge discovery. In this paper, we build a multiagent system that can track the social correlations within a group of people based on video data. Our information diffusion system targets small groups of people, possibly composed of office workers, meeting attendees, etc. Adopting a multiagent architecture to study the influential correlations in a social network is an adequate choice since it maintains the scalability and robustness of the system. The correlation amongst the nodes of the social network is built on the basis of facial emotions. We evaluated the method in a social network with scripted discussions. Our results show that the emotion propagation was effectively reflected in the predicted social influence correlation.

This paper presents a dramatic model for self-regulation of social exchange processes in multiagent systems, based on the concepts of Drama Theory. The model has five phases of dramatic resolution, which involve feelings, emotions, trust and reputation. Agents with different social exchange strategies interact each other in order to maximize their strategy-based fitness functions. The objective is to obtain a more natural model than the ones existing in the literature, which are based on (partially observable) Markov decision processes or in game theory, so that it can be applied in real-world applications. We aim at promoting more balanced and fair multiagent interactions, increasing the number of successful social exchanges and, thus, promoting the continuity of social exchanges.

Over the last few years educators have increasingly incorporated game simulations into higher education computer science curricula. Experiences have proved that students respond enthusiastically to these courses. According to this, this paper presents an evolved version of the JGOMAS simulator, which is a simulation game where students design and implement different types of agents and strategies in order to win the game. This new version allows students to practice different technologies related to the multi-agent paradigm as coordination, cooperation or decision-making.

Public Transportation Systems (PTSs) are always subjected to disturbances and need a real time monitoring and control to maintain its performance at acceptable levels. In PTS, several types of disturbances can affect buses such as accidents, delays and traffic jams that can also affect schedules so dramatically that these schedules could become useless. Consequently, it becomes a necessity to develop a Decision Support System (DSS) able to help human regulator in managing PTS efficiently, and to provide users with high quality services, in terms of punctuality, frequency and productivity. In this paper, a reactive and decentralized DSS is developed for the control of PTS based on the biological immune theory. This DSS is an artificial immune system, which presents many interesting capabilities, including identification, learning, memory and distributed parallel processing. Through experimental validation, we show that this exploratory approach seems to be promising.

One of the main objectives of AI is modelling human reasoning. Since preference information is an indispensable component of common-sense reasoning, the two should be studied in tandem. Argumentation is an established branch of AI dedicated to this task. In this paper, we study how argumentation with preferences models human intuition behind a particular decision making scenario concerning reasoning with rules and preferences. To this end, we present an example of a common-sense reasoning problem complemented with a survey of decisions made by human respondents. The survey reveals an answer that contrasts with solutions offered by various argumentation formalisms. We argue that our results call for advancements of approaches to argumentation with preferences as well as for examination of the type of problems of reasoning with preferences put forward in this paper. Our work contributes to the line of research on preference handling in argumentation, and it also enriches the discussions on the increasingly important topic of preference treatment in AI at large.

A bargaining problem is a cooperative game in which players are permitted to negotiate before the game is played. Bargaining theory can be used to economic interactions such as union negotiations, international trade agreements, and duopolies. A general theory of bargaining games thus has a wide application to many areas of economics and political science.

In this research work we propose a new approach to estimate the number of passengers in a public transportation or determinate the users’ route path based on probe requests of users mobile device through collected data in wireless access point. This data is manipulated to extract the information about the numbers of users with mobile devices and track their route path and time.

In the last two decades, energy planning decision-making (EPDM), especially the evaluation and prioritization of renewable energy sources (RES), has attracted significant attention. The decision-making process is aligned with several sources that can be uncertain, including incomplete information, limited domain knowledge from decision-makers, and failures to provide accurate judgments from experts. In this study, the Evidential Reasoning (ER) approach is developed to manage the expanding complexities and uncertainties in assessment problems. The ER approach is employed as a multiple criteria framework to assess the appropriateness regarding the use of different renewable energy technologies. A case study is provided to illustrate the implementation process. Results show that using the ER approach when assessing the sustainability of different RES under uncertainty allows providing robust decisions, which brings out a more accurate, effective, and better-informed EPDM tool to conduct the evaluation process.

Assistive technologies help users with disabilities (physical, sensory, intellectual) to perform tasks that were difficult or impossible to execute. Thus, the user autonomy is increased through this technology. Although some adaptation of the user might be needed, the effort should be minimum in order to use devices that convey assistive functionalities. In cognitive disabilities a common diminished capacity is orientation, which is crucial for the autonomy of an individual. There are several research works that tackle this problem, however they are essentially concerned with user guidance and application interface (display of information). The work presented herein aims to overcome these systems through a framework of Speculative Computation, which adds a prediction feature for the next move of the user. With an anticipation feature and a trajectory mining module the user is guided through a preferred path receiving anticipated alerts before a possible shift in the wrong direction.

Supporting group decision-making is a complex process, especially when decision-makers have no opportunity to gather at the same place and at the same time. Besides that, finding solutions may be difficult in case representing agents are not able to understand the process and support the decision-maker accordingly. Here we propose a model and an algorithm that will allow the agent to analyse tendencies. This way we intend that agents can achieve decisions with more quality and with higher levels of consensus. Our model allows the agent to redefine his objectives to maximize both his and group satisfaction. Our model proved that agents that use it will obtain higher average levels of consensus and satisfaction. Besides that, agents using this model will obtain those higher levels of consensus and satisfaction in most of the times compared to agents that do not use it.

In this paper we show how multiple BDI agents, enhanced with temporal difference learning capabilities, learn their utility function, while they are concurrently exploring an uncertain environment. We focus on the programming aspects of the agents using the Jason agent-oriented programming language. We also provide experimental results showing the behavior of multiple agents acting in a Markovian grid environment. We consider agents with the perception function affected by the intermittent faults and Gaussian noise, as well as agents for which their action function is not always successful.

In this paper, we describe a technique to develop simulated user agents that are able to interact with dialog systems. By means of these agents, it is possible not only to automatically evaluate the overall operation of the dialog system, but also to assess the impact of the user responses on the decisions that are selected by the system. The selection of the user responses by the simulated user agent are based on a statistical model that is automatically learned from a dialog corpus. The complete history of the interaction is considered to carry out this selection. The paper describes the application of this technique to evaluate a practical dialog system providing tourist information and services.

Ambient Assisted Living (AAL) is an emergent area that provides useful mechanisms that allows tracking elders through sensoring. For AAL systems, it is very important to provide information fusion techniques, which merge the information available in sensors available in different devices like the smartphones to infer possible risk situations for elders in outdoor environments. The Data Fusion Model is the most widely used method for categorizing data fusion-related functions. In previous works we have developed SafeRoute, an AAL system that pretends monitoring elders in their day-to-day daily living activities in outdoor environments. In this context, this paper presents a specific proposal of application of the JDL Data Fusion Model to tracking old persons in outdoor environments. We additionally present the social interaction model in the context of the SafeRoute system, showing the interactions between caregivers and elders and including new contextual elements to make more efficient the tracking process.

This research uses data mining techniques to establish data predicting consumption of products that are grown in areas of the Andean region of Ecuador by relatives and marketed in alternative circuits that prevent intermediary called Cialcos to improve their income groups.

Internet upswing has entailed a structural change for journalism in general and the press in particular. The emergence of a new horizontal, low cost and accessible space for communication, has brought profound changes in journalism, both on the production and distribution. In this paper, we present a novel agent-based social platform which aims to improve the organization, management and distribution of the media contents through the application of artificial intelligence techniques.

This paper asserts that a multi-perspective viewpoint must be taken in the design of a computational system support capability for decision-making. We offer views from a Decision-Science slant, a Systemic Architectural view, and the need for technological support to realize improvements in analytical rigor. We have been researching and evolving the design of an analysis tool framework exploiting the hybrid concepts of a Belief-based Argumentation and Story-based subsystem. The notion of rigor, defined as a quality measure on the reasoning/analysis process, is one overarching principle of our approach, driven by the need for the associated analysis/decision-support product quality that complex modern problems demand. Our approach to the design of a mixed-initiative analysis tool is highly multidisciplinary and has taken account of an exhaustive review of the relevant literature along each viewpoint.

A steadily growing pervasion of the energy grid with communication technology is widely seen as an enabler for new computational coordination techniques for renewable, distributed generation as well as for controllable consumers. One important task is the ability to group together in order to jointly gain enough suitable flexibility and capacity to assume responsibility for a specific control task in the grid. We present a fully decentralized coalition formation approach based on an established heuristic for predictive scheduling with the additional advantage of keeping all information about local decision base and local operational constraints private. The approach is evaluated in several simulation scenarios with different type of established models for integrating distributed energy resources.

Due to politically defined goals to raise the share of renewable energy, the landscape of electricity production has changed in recent years. Normally, a decision to invest in new generation capacity by generation companies is often based on profit maximization criteria. Criteria considering the costs resulting from the required expansion or construction of new transmission capacity are only playing a minor role, if any. This paper introduces an integrated model based on a multi-agent system to simulate the investment and decision behavior of the relevant entities in the liberalized energy market and their impact on social welfare. The interaction between the modelled market entities is based on a non-cooperative game theoretic approach. Its functionality is demonstrated within a small application example.

As a result of fast growing share of renewable energy production in the energy market the management of power and its distribution becomes more and more complex. The here presented Energy Option Model (EOM) seems to be a promising solution to handle this newly arisen complexity. This paper will present the EOM and analyze its capabilities in centralized evaluation of aggregated systems. The example use-case will be the charging process of a fleet of electric vehicles. While the results support the potential of the EOM to implement coordination strategies for aggregations of systems, they also show the general limitations of centralized control solutions for larger groups of systems in the context of smart grids.

The actual flexibility of the electricity sector, with a distributed nature and new players, such as the smart grid operator and several types of aggregators, brings new business models and introduces new challenges from the power systems technical operation point of view. In this context, the Network Operator Agent of the Multi-Agent Simulator of Competitive Electricity Markets (MASCEM) plays a crucial role, not only in the scope of the technical validation of the economic transactions established by the market, but also has an agent that can be supporting the grid operation under the scope of a smart grid. A set of new features has been added to the Network Operator making it a “new agent”, bringing a more effective decision support, from the grid technical operation point of view, and achieving its usefulness beyond MASCEM. In this paper the new features are described. A case study is also included to better illustrate the approach and to highlight its usefulness under the scope of a smart grid scenario.

Power systems worldwide are complex and challenging environments. The increasing necessity for an adequate integration of renewable energy sources is resulting in a rising complexity in power systems operation. Multi-agent based simulation platforms have proven to be a good option to study the several issues related to these systems, including the involved players that act in this domain. To take better advantage of these systems, their integration is mandatory. The main contribution of this paper is the development of the Electricity Markets Ontology, which integrates the essential concepts necessary to interpret all the available information related to electricity markets, while enabling an easier cooperation and adequate communication between related systems. Additionally, the concepts and rules defined by this ontology can be extended and complemented according to the needs of other simulation and real systems in this area. Each system’s particular ontology must import the proposed ontology, thus enabling the effective interoperability between independent systems.

Computer Supported Collaborative Learning (CSCL) is a computational approach allowing students−being organized into groups− to work together for a common goal and discuss from different points of view always seeking to improve learning processes. The context-awareness concept, which is inherent to humans when performing any learning activity, becomes the main component for monitoring activities in virtual learning environments. The aim of this paper is to integrate context-awareness services to enhance CSCL environments supported by Multi-Agent Systems. The awareness-group-agent that composes the system architecture provides the coordination of several agents responsible of handling each of the 11 context-awareness services proposed. This characteristic allows both students and teachers at a given time be aware of their teamwork progress status during execution of the CSCL-MAS environment. In order to validate the incorporation of context-aware services a prototype was built and tested through a case study. Results obtained demonstrate the effectiveness of using this kind of approaches in collaborative learning environments which constitutes an attempt to improve learning processes.

In a face-class, where the student group is heterogeneous, it is necessary to select the most appropriate educational resources that support learning for all. In this sense, multi-agent system (MAS) can be used to simulate the features of the students in the group, including their learning style, in order to help the professor find the best resources for your class. In this paper, we present MAS to recommendation educational resources for group students, simulating their profiles and selecting resources that best fit. Obtained promising results show that proposed MAS is able to delivered educational resources for a student group.

This paper presents the implementation and evaluation of ILOMAS, an architectural model designed to select Learning Objects (LO) for e-learning, based on multi-agent systems. The proposed model extends the Intelligent Learning Objects approach through the use of a BDI agent architecture, allowing the communication with the instructional resources that constitute the LO according to the SCORM standard. A prototype implementation is presented to evaluate the proposed model.

The paper proposes a model for the continuous improvement of academic teaching aimed at delivering a programmed excellent learning in perspective. The proposed Teaching Evaluation Model (TEM) is a dynamic and open system based on the Deming Cycle (PLAN-DO-CHECK-ACT). The objective pursued by the model is to match the expected learning with real learning. The results of this work are intended to highlight the field application of the TEM approach. Specifically, the application consists of a web-based tool conceived and designed to allow teachers and institutions to build a continuous improvement of the teaching and learning processes. By enhancing the interior design education and hence the profession itself, the model reveals that technology-enhanced assessment may deliver tangible benefits for learners, teachers and institutions.