About 25 years ago, the Nobel laureate Herbert A. Simon and other top Management Science/Operatio... more About 25 years ago, the Nobel laureate Herbert A. Simon and other top Management Science/Operations Research (MS/OR) and Artificial Intelligence (AI) researchers, suggested that an integration of the two disciplines would improve the design of decision-making support tools in organizations. The suggested integrated system has been called an intelligent decision-making support system (i-DMSS). In this chapter, we use an existing conceptual framework posed to assess the capabilities and limitations of the i-DMSS concept, and through a conceptual metaanalysis research of the Decision Support System (DSS) and AI literature from 1980 to 2004, we develop a strategic assessment of the initial proposal. Such an analysis reveals support gaps that suggest further development of the initial i-DMSS concept is needed. We offer recommendations for making the indicated improvements in i-DMSS design, development, and application.
Design and evaluation frameworks for Decision-making Support Systems (DMSS) and Intelligent DMSS ... more Design and evaluation frameworks for Decision-making Support Systems (DMSS) and Intelligent DMSS (i-DMSS) have been posed in last 20 years. Useful findings to match the required general system's capabilities with decision phases and steps in several managerial levels have been also generated. However, current status of i-DMSS capabilities suggests that the full realization of them is still a long-term aim. This paper seeks to advance in it. By extending a previous work and integrating the Chandrasekaran's task-structure concept ...
Standards and models of processes–such as ISO/IEC 9000 standard for deploying quality management ... more Standards and models of processes–such as ISO/IEC 9000 standard for deploying quality management systems–have been developed for international organizations to promote the utilization of best managerial and engineering practices. However, given their conceptual density, a large number of concepts, composite concepts, and interrelationships, understanding them cannot be considered a trivial cognitive task for new readers and decision makers. Consequently, some IT-supported systems have been used to improve ...
We present our bayesian-modeler agent which uses a probabilistic approach for agent modeling. It ... more We present our bayesian-modeler agent which uses a probabilistic approach for agent modeling. It learns models about the others using a bayesian mechanism and then it plays in a rational way using a decision-theoretic approach. We also describe our empirical study on evaluating the competitive advantage of our modeler agent. We explore a range of strategies from the least- to most-informed one in order to evaluate the lower- and upper-limits of a modeler agent’s performance. For comparison purposes, we also developed and experimented with other different modeler agents using reinforcement learning techniques. Our experimental results showed how an agent that learns models about the others, using our probabilistic approach, reach almost the optimal performance of the oracle agent. Our experiments have also shown that a modeler agent using a reinforcement learning technique have a performance not as good as the bayesian modeler’ performance. However, it could be competitive under different assumptions and restrictions.
Knowledge and Information distribution is indeed one of the main processes in Knowledge Managemen... more Knowledge and Information distribution is indeed one of the main processes in Knowledge Management. Today, most Information Technology tools for supporting this distribution are based on repositories accessed through Web-based systems. This approach has, however, many practical limitations, mainly due to the strain they put on the user, who is responsible of accessing the right Knowledge and Information at the right moments. As a solution for this problem, we have proposed an alternative approach which is based on the notion of delegation of distribution tasks to synthetic agents, which become responsible of taking care of the organization's as well as the individuals' interests. In this way, many Knowledge and Information distribution tasks can be performed on the background, and the agents can recognize relevant events as triggers for distributing the right information to the right users at the right time. In this paper, we present the JITIK approach to model knowledge and information distribution, giving a high-level account of the research made around this project, emphasizing two particular aspects: a sophisticated argument-based mechanism for deciding among conflicting distribution policies, and the embedding of JITIK agents in enterprises using the service-oriented architecture paradigm. It must be remarked that a JITIK-based application is currently being implemented for one of the leading industries in Mexico.
Communication among agents in swarm intelligent systems and more generally in multiagent systems,... more Communication among agents in swarm intelligent systems and more generally in multiagent systems, is crucial in order to coordinate agents’ activities so that a particular goal at the collective level is achieved. From an agent’s perspective, the problem consists in establishing communication policies that determine what, when, and how to communicate with others. In general, communication policies will depend on the nature of the problem being solved. This means that the solvability of problems by swarm intelligent systems depends, among other things, on the agents’ communication policies, and setting an incorrect set of policies into the agents may result in finding poor solutions or even in the unsolvability of problems. As a case study, this paper focus on the effects of letting agents use different communication policies in ant-based clustering algorithms. Our results show the effects of using different communication policies on the final outcome of these algorithms.
The present document explores how air pollution can be assessed from a multiagent point of view. ... more The present document explores how air pollution can be assessed from a multiagent point of view. In order to do so, a traffic system was simulated using agents as a way to measure if air pollution levels go down when the traffic lights employ a multigent cooperative system that negotiates the green light duration of each traffic light, in order to minimize the time a car has to wait to be served in an intersection. The findings after running some experiments where lanes of each direction are congested incrementally showed, that using this technique, there is a significant decrease in air pollution over the simulated area which means that traffic lights controlled by the multiagent system do improve the levels of air pollution.
We propose the use of a Fuzzy Naive Bayes classifier with a MAP rule as a decision making module ... more We propose the use of a Fuzzy Naive Bayes classifier with a MAP rule as a decision making module for the RoboCup Soccer Simulation 3D domain. The Naive Bayes classifier has proven to be effective in a wide range of applications, in spite of the fact that the conditional independence assumption is not met in most cases. In the Naive Bayes classifier, each variable has a finite number of values, but in the RoboCup domain, we must deal with continuous variables. To overcome this issue, we use a fuzzy extension known as the Fuzzy Naive Bayes classifier that generalizes the meaning of an attribute so it does not have exactly one value, but a set of values to a certain degree of truth. We implemented this classifier in a 3D team so an agent could obtain the probabilities of success of the possible action courses given a situation in the field and decide the best action to execute. Specifically, we use the pass evaluation skill as a test bed. The classifier is trained in a scenario where there is one passer, one teammate and one opponent that tries to intercept the ball. We show the performance of the classifier in a test scenario with four opponents and three teammates. After a brief introduction, we present the specific characteristics of our training and test scenarios. Finally, results of our experiments are shown.
Learning and making decisions in a complex uncertain multiagent environment like RoboCup Soccer S... more Learning and making decisions in a complex uncertain multiagent environment like RoboCup Soccer Simulation 3D is a non-trivial task. In this paper, a probabilistic approach to handle such uncertainty in RoboCup 3D is proposed, specifically a Naive Bayes classifier. Although its conditional independence assumption is not always accomplished, it has proved to be successful in a whole range of applications. Typically, the Naive Bayes model assumes discrete attributes, but in RoboCup 3D the attributes are continuous. In literature, Naive Bayes has been adapted to handle continuous attributes mainly using Gaussian distributions or discretizing the domain, both of which present certain disadvantages. In the former, the probability density of attributes is not always well-fitted by a normal distribution. In the latter, there can be loss of information. Instead of discretizing, the use of a Fuzzy Naive Bayes classifier is proposed in which attributes do not take a single value, but a set of values with a certain membership degree. Gaussian and Fuzzy Naive Bayes classifiers are implemented for the pass evaluation skill of 3D agents. The classifiers are trained with different number of training examples and different number of attributes. Each generated classifier is tested in a scenario with three teammates and four opponents. Additionally, Gaussian and Fuzzy approaches are compared versus a random pass selector. Finally, it is shown that the Fuzzy Naive Bayes approach offers very promising results in the RoboCup 3D domain.
About 25 years ago, the Nobel laureate Herbert A. Simon and other top Management Science/Operatio... more About 25 years ago, the Nobel laureate Herbert A. Simon and other top Management Science/Operations Research (MS/OR) and Artificial Intelligence (AI) researchers, suggested that an integration of the two disciplines would improve the design of decision-making support tools in organizations. The suggested integrated system has been called an intelligent decision-making support system (i-DMSS). In this chapter, we use an existing conceptual framework posed to assess the capabilities and limitations of the i-DMSS concept, and through a conceptual metaanalysis research of the Decision Support System (DSS) and AI literature from 1980 to 2004, we develop a strategic assessment of the initial proposal. Such an analysis reveals support gaps that suggest further development of the initial i-DMSS concept is needed. We offer recommendations for making the indicated improvements in i-DMSS design, development, and application.
Design and evaluation frameworks for Decision-making Support Systems (DMSS) and Intelligent DMSS ... more Design and evaluation frameworks for Decision-making Support Systems (DMSS) and Intelligent DMSS (i-DMSS) have been posed in last 20 years. Useful findings to match the required general system's capabilities with decision phases and steps in several managerial levels have been also generated. However, current status of i-DMSS capabilities suggests that the full realization of them is still a long-term aim. This paper seeks to advance in it. By extending a previous work and integrating the Chandrasekaran's task-structure concept ...
Standards and models of processes–such as ISO/IEC 9000 standard for deploying quality management ... more Standards and models of processes–such as ISO/IEC 9000 standard for deploying quality management systems–have been developed for international organizations to promote the utilization of best managerial and engineering practices. However, given their conceptual density, a large number of concepts, composite concepts, and interrelationships, understanding them cannot be considered a trivial cognitive task for new readers and decision makers. Consequently, some IT-supported systems have been used to improve ...
We present our bayesian-modeler agent which uses a probabilistic approach for agent modeling. It ... more We present our bayesian-modeler agent which uses a probabilistic approach for agent modeling. It learns models about the others using a bayesian mechanism and then it plays in a rational way using a decision-theoretic approach. We also describe our empirical study on evaluating the competitive advantage of our modeler agent. We explore a range of strategies from the least- to most-informed one in order to evaluate the lower- and upper-limits of a modeler agent’s performance. For comparison purposes, we also developed and experimented with other different modeler agents using reinforcement learning techniques. Our experimental results showed how an agent that learns models about the others, using our probabilistic approach, reach almost the optimal performance of the oracle agent. Our experiments have also shown that a modeler agent using a reinforcement learning technique have a performance not as good as the bayesian modeler’ performance. However, it could be competitive under different assumptions and restrictions.
Knowledge and Information distribution is indeed one of the main processes in Knowledge Managemen... more Knowledge and Information distribution is indeed one of the main processes in Knowledge Management. Today, most Information Technology tools for supporting this distribution are based on repositories accessed through Web-based systems. This approach has, however, many practical limitations, mainly due to the strain they put on the user, who is responsible of accessing the right Knowledge and Information at the right moments. As a solution for this problem, we have proposed an alternative approach which is based on the notion of delegation of distribution tasks to synthetic agents, which become responsible of taking care of the organization's as well as the individuals' interests. In this way, many Knowledge and Information distribution tasks can be performed on the background, and the agents can recognize relevant events as triggers for distributing the right information to the right users at the right time. In this paper, we present the JITIK approach to model knowledge and information distribution, giving a high-level account of the research made around this project, emphasizing two particular aspects: a sophisticated argument-based mechanism for deciding among conflicting distribution policies, and the embedding of JITIK agents in enterprises using the service-oriented architecture paradigm. It must be remarked that a JITIK-based application is currently being implemented for one of the leading industries in Mexico.
Communication among agents in swarm intelligent systems and more generally in multiagent systems,... more Communication among agents in swarm intelligent systems and more generally in multiagent systems, is crucial in order to coordinate agents’ activities so that a particular goal at the collective level is achieved. From an agent’s perspective, the problem consists in establishing communication policies that determine what, when, and how to communicate with others. In general, communication policies will depend on the nature of the problem being solved. This means that the solvability of problems by swarm intelligent systems depends, among other things, on the agents’ communication policies, and setting an incorrect set of policies into the agents may result in finding poor solutions or even in the unsolvability of problems. As a case study, this paper focus on the effects of letting agents use different communication policies in ant-based clustering algorithms. Our results show the effects of using different communication policies on the final outcome of these algorithms.
The present document explores how air pollution can be assessed from a multiagent point of view. ... more The present document explores how air pollution can be assessed from a multiagent point of view. In order to do so, a traffic system was simulated using agents as a way to measure if air pollution levels go down when the traffic lights employ a multigent cooperative system that negotiates the green light duration of each traffic light, in order to minimize the time a car has to wait to be served in an intersection. The findings after running some experiments where lanes of each direction are congested incrementally showed, that using this technique, there is a significant decrease in air pollution over the simulated area which means that traffic lights controlled by the multiagent system do improve the levels of air pollution.
We propose the use of a Fuzzy Naive Bayes classifier with a MAP rule as a decision making module ... more We propose the use of a Fuzzy Naive Bayes classifier with a MAP rule as a decision making module for the RoboCup Soccer Simulation 3D domain. The Naive Bayes classifier has proven to be effective in a wide range of applications, in spite of the fact that the conditional independence assumption is not met in most cases. In the Naive Bayes classifier, each variable has a finite number of values, but in the RoboCup domain, we must deal with continuous variables. To overcome this issue, we use a fuzzy extension known as the Fuzzy Naive Bayes classifier that generalizes the meaning of an attribute so it does not have exactly one value, but a set of values to a certain degree of truth. We implemented this classifier in a 3D team so an agent could obtain the probabilities of success of the possible action courses given a situation in the field and decide the best action to execute. Specifically, we use the pass evaluation skill as a test bed. The classifier is trained in a scenario where there is one passer, one teammate and one opponent that tries to intercept the ball. We show the performance of the classifier in a test scenario with four opponents and three teammates. After a brief introduction, we present the specific characteristics of our training and test scenarios. Finally, results of our experiments are shown.
Learning and making decisions in a complex uncertain multiagent environment like RoboCup Soccer S... more Learning and making decisions in a complex uncertain multiagent environment like RoboCup Soccer Simulation 3D is a non-trivial task. In this paper, a probabilistic approach to handle such uncertainty in RoboCup 3D is proposed, specifically a Naive Bayes classifier. Although its conditional independence assumption is not always accomplished, it has proved to be successful in a whole range of applications. Typically, the Naive Bayes model assumes discrete attributes, but in RoboCup 3D the attributes are continuous. In literature, Naive Bayes has been adapted to handle continuous attributes mainly using Gaussian distributions or discretizing the domain, both of which present certain disadvantages. In the former, the probability density of attributes is not always well-fitted by a normal distribution. In the latter, there can be loss of information. Instead of discretizing, the use of a Fuzzy Naive Bayes classifier is proposed in which attributes do not take a single value, but a set of values with a certain membership degree. Gaussian and Fuzzy Naive Bayes classifiers are implemented for the pass evaluation skill of 3D agents. The classifiers are trained with different number of training examples and different number of attributes. Each generated classifier is tested in a scenario with three teammates and four opponents. Additionally, Gaussian and Fuzzy approaches are compared versus a random pass selector. Finally, it is shown that the Fuzzy Naive Bayes approach offers very promising results in the RoboCup 3D domain.
Uploads
Papers by leonardo garrido