Citation Context

... the opportunity to exploit the dynamics of type change to improve its own rewards over time. That means the agents have another new dimension besides the basic exploration vs. exploitation tradeoff (=-=Kaelbling et al., 1996-=-) common in typical reinforcement learning scenarios to improve its own rewards over time. Notice that, the change in the types of the agents do occur in a variety of coalition formation environments ...

Citation Context

...sk to maximize its rewards over time with respect to its type. In our design and implementation of ADLIT, we formulate the type change of the agents as a Partially Observable Markov Decision Process (=-=Kaelbling et al., 1998-=-) and design a coalition formation protocol and approximation strategies. The coalition formation protocol specifies how a set of agents coordinate among themselves to pick (1) the types of the member...

Citation Context

...t increases the effective working memory capacity allowing acceleration of future learning processes. Finally, according to the recent research work on perceptual and motor acquisitions (Fitts, 1964; =-=Anderson, 1982-=-) the pace of skill acquisition for a learner accelerates in the beginning and slows down to a stable state, leading to: Observation 5: A student’s available working memory for a topic is proportional...

Citation Context

...reward gain in addition to considering current task rewards. Finally, as an improvement over our One-Step Lookahead, in the Learning by Exploration (LEA) method, the agents use a simulated annealing (=-=Russel and Norvig, 2003-=-, p. 115) approach to improve their choice of learning actions by balancing its exploration and exploitation. 1.3 Contributions Our research work contributes to the advancements of the solution techni...

Citation Context

... in the environment and (2) the inherent inaccuracy in measuring students’ attributes that define their types - i.e., the inaccuracy in the measurement of students’ attributes using tests or surveys (=-=Cortina, 1993-=-). Also, the participating students’ types (e.g., knowledge and ability to solve problems) change due to their interactions with their coalition members. Analyzing the state of the art of MCFP-M type ...

Citation Context

...ice that each mediator agent is able to observe different human user models (i.e., the model of their assigned human users) simultaneously and thus together divide the task (cf. division learning in (=-=Dillenbourg, 1999-=-) of observing the entire set of possible model47 Figure 3.4: Update of Bayesian network output combinations. This learning strategy is designed to allow the agents to learn the dynamics of the enviro...

Citation Context

...the agents’ ability to effectively use the type change to improve their utility over time. Although the exact value of an environment parameter is not important in multiagent simulation environments (=-=Bonabeau, 2002-=-) choosing a very low value would mean that for the same input values of Equation 25, the agents would observe different outcomes. Thus, 141 those low values would make it difficult for the agents to ...

Citation Context

...et student groups had higher standard deviation on average than the treatment set students. Homogeneous student groups have been suggested by the collaborative learning researchers and practitioners (=-=Johnson and Johnson, 1999-=-) as a strategy to form student groups especially for settings where students on average have high expertise on the task they are collaborating to solve. Being an advanced course in computer science, ...

Citation Context

...o do a task, watches as the apprentice practices portions of the task, and then turns over more and more responsibility until the apprentice is proficient enough to accomplish the task independently (=-=Collins et al., 1991-=-). • Learning by Practice. This type of learning occurs when a human user applies his or her existing knowledge to solve an assigned problem. This type of learning is very common in situations where e...

Citation Context

...CFP-A problem are described in Table 2.3. Table 2.3: Agent-Acted Coalition Formation Research I (Research) Solved Problem (Shehory and Kraus, 1995) Coalition formation as a method of task allocation (=-=Sandholm and Lesser, 1997-=-) Coalition formation for computationally bounded agents (Dieckmann and Schwalbe, 1998) Coalition formation in dynamic environments (Sandholm et al., 1999) Computational complexity of coalition format...

Citation Context

... the task solution quality dynamics or the agent type change dynamics (Section 6.4.5). So, we plan to change the coalition formation protocol so that an agent can use multiagent cooperative learning (=-=Panait and Luke, 2005-=-) to exchange their learned probability values (regarding the environment dynamics) and learn from one another. Such cooperative learning would enable the agents to improve their estimations more effi...

Citation Context

...ained from collaborative knowledge of his or her group members through interaction. That collaborative knowledge is in turn produced by individual knowledge of the interacting group members. Kreijns (=-=Kreijns et al., 2003-=-) describe the interaction between students as the key to collaboration among group members. Furthermore, researchers (der Linden and Renshaw, 2001) suggest that collaborative learning occurs from the...

Citation Context

...udent and his or her peer (as perceived by the student) change according to the frequency, extent, and quality of collaboration (e.g., how many times did my peer helped me). 4.1.4 Scaffolding Bruner (=-=Bruner, 1985-=-) and Cazden (Cazden, 1983) define scaffolding as the act of providing assistance to a child so that he or she is able to carry out a task (e.g., solve a problem) that he or she cannot do by herself. ...

Citation Context

... increases, it increases the effective working memory capacity allowing acceleration of future learning processes. Finally, according to the recent research work on perceptual and motor acquisitions (=-=Fitts, 1964-=-; Anderson, 1982) the pace of skill acquisition for a learner accelerates in the beginning and slows down to a stable state, leading to: Observation 5: A student’s available working memory for a topic...

Citation Context

...us only on the coalition formation- and scaffolding-related aspects of the agent-mediator cases of the ADLIT. A research approach that investigates the issue of adjustable autonomy could be found in (=-=Screrri et al., 2002-=-; Maheswaran et al., 2004). 35 stochastic–meaning the behavior of the human human users cannot be accurately modeled; only modeled with probability. Assumption 7 (Reward) The rewards achievable by the...

Citation Context

...ncy denotes the reward-to-effort ratio incurred by the agents who solve a task. Multiagent coalition formation techniques can be used to solve a variety of real-world problems, e.g., task allocation (=-=Shehory and Kraus, 1995-=-), bargain buying in business transactions (Yamamoto and Sycara, 2001), and forming student groups in computer-supported collaborative learning environments (Vassileva et al., 2003). In our research, ...

Citation Context

... those problems is that the state and action spaces are continuous. As a result, designing and implementing a multiagent reinforcement learning environment becomes difficult and sometimes impossible (=-=Busoniu et al., 2008-=-). Furthermore, real-world human coalition formation contains incomplete and uncertain information which complicates the learning problem for the agents further. In our research work, we tackle the le...

Citation Context

...ate the type change of the agents due to the implicit and explicit learning in the repeated coalition formation environment of ADLIT as a partially observable MDP (POMDP) (a.k.a. a belief state MDP) (=-=Dearden et al., 1999-=-). Let us assume that the agents i ∈ N have initial belief Bi that contains the estimations regarding the dynamics and the openness of the environment. As an agent joins coalitions, observes the resul...

Citation Context

...he coalition that provides the maximal current-task reward without taking advantage of the typechanging impact of the implicit and explicit learning actions. In the One-Step Lookahead (OSL) strategy (=-=Emery-Montemerlo et al., 2004-=-), the agents look ahead into the immediate round of coalition formation to approximate the reward gain in the future. Finally, in the Learning by Exploration (LEA) strategy, the agents utilize the ex...

Citation Context

...n techniques to solve important but yet to be automatically solved real-world coalition formation prob8lems e.g., businesses forming coalitions to improve their earnings or to solve difficult tasks ((=-=Khanna and Yafeh, 2007-=-); (Iacobucci and Rosa, 2005)) and political coalition formation to pass legislature (Kreppel and Tsebelis, 1999)) well. 1.2 Solution Approach To solve MCFP-M, we propose the Integrated Human Coalitio...

Citation Context

...out the task they are solving. • Learning by Reflection/Self-Expression. This type of learning occurs when a human user rethinks his or her own solution and analyzes his or her self-thinking process (=-=Schon, 1987-=-, p. 28). • Learning by Apprenticeship. In learning by apprenticeship, the expert teacher shows the apprentice (student) how to do a task, watches as the apprentice practices portions of the task, and...

Citation Context

...dentifies how each solution approach addresses the critical aspects (ca1, ca2 in Section 2.1) of MCFP-A. Table 2.4: Agent-Acted Coalition Formation Research II (Research): Problem, Solution. ca1 ca2 (=-=Ketchpel, 1994-=-): Coalition formation for tasks that require heterogeneous agent expertise and where the agents are uncertain about the value of a coalition. An auction-based coalition formation algorithm is propose...

Citation Context

...ons (Breban and Vassileva, 2002) Coalition formation in business environments (coalitions between customers and vendors) (Chalkiadakis and Boutilier, 2003) Coalition formation for agent coordination (=-=Dang and Jennings, 2004-=-) Optimization of social payoff in coalition formation 26 (Research) Solved Problem (Abdallah and Lesser, 2004) Coalition formation as a method of agent organization (Yokoo et al., 2005) Solution prop...

Citation Context

... (2006); Chalkiadakis (2007) do not adequately exploit the agents’ type-changing influences into account in their framework or solution approach. For example, the agents in the solutions proposed in (=-=Kraus et al., 2003-=-); (Chalkiadakis and Boutilier, 2004); (Hosam and Khaldoun, 2006); (Chalkiadakis, 2007) do not consider the impact of an agent influencing type changes in its coalition members by learning actions. Du...

Citation Context

...lankenburg and Klusch, 2004): Forming coalition of the most reliable agents in repeated coalitional games. A trusted kernel-based, encryptionenabled coalition formation protocol was designed. Yes na (=-=Gaston and DesJardins, 2005-=-): Finding effective and efficient social network in dynamic environment. Distributed coalition formation algorithm for the agents was provided. na na (Chalkiadakis, 2007; Chalkiadakis and Boutilier, ...

Citation Context

...oup formation, we have boot102 strapped the Bayesian network to form heterogeneous groups containing students of a variety of levels of competence, i.e., an introductory classroom, as recommended in (=-=Cress and Kimmerle, 2008-=-). The analysis of our results shows that, the average standard deviation of the treatment students’ total prior scores were higher (13.9 vs. 8.3 with p < 0.05) than the control set students. This imp...

Citation Context

...assroom. - MHCF: To improve upon the design of the VALCAM algorithm, we have designed the MHCF algorithm that uses a negotiation-based multiagent coalition formation algorithm and a Bayesian network (=-=Cowell, 1998-=-, pp. 9-27) to learn the composition of student groups that would allow the members to solve the current task well as well as improve their behavior to earn higher rewards for future tasks. To investi...

Citation Context

...ure has mostly focused on how agents learn from each other when they interact or negotiate or collaborate (Chalkiadakis, 2007; Kraus et al., 2003; Chan and Le14 ung, 2008; Tsvetovat and Sycara, 2000; =-=Klusch and Gerber, 2002-=-; Kraus et al., 2004) but not on what learning can do to an agent’s type when forming or joining a coalition, and parlaying that into the future. Our solution, on the other hand, integrally considers ...

Citation Context

...(skills, knowledge, etc.) can change because of their learning interactions with other agents in the environment. Examples of this type of coalition formation can be found in electronic marketplaces (=-=Lerman and Shehory, 2000-=-), trust-based coalition formations (Griffiths and Luck, 2003), etc. All of the distinct characteristics of MCFP-M also exist in these coalition formation problems but with the difference that the age...

Citation Context

... complicated missions. A negotiation-based distributed coalition formation based on the agents’ social knowledge of acquaintance models was provided. na na (Vassileva et al., 1999; Bull et al., 2001; =-=Greer et al., 2001-=-; Vassileva et al., 2003): Learner coalition formation for providing 1-to-1 peer help in collaborative learning environment. A framework where agents watch and develop probabilistic models of the user...

Citation Context

...(2007) do not adequately exploit the agents’ type-changing influences into account in their framework or solution approach. For example, the agents in the solutions proposed in (Kraus et al., 2003); (=-=Chalkiadakis and Boutilier, 2004-=-); (Hosam and Khaldoun, 2006); (Chalkiadakis, 2007) do not consider the impact of an agent influencing type changes in its coalition members by learning actions. Due to the necessary collaboration of ...

Citation Context

...g interactions with other agents in the environment. Examples of this type of coalition formation can be found in electronic marketplaces (Lerman and Shehory, 2000), trust-based coalition formations (=-=Griffiths and Luck, 2003-=-), etc. All of the distinct characteristics of MCFP-M also exist in these coalition formation problems but with the difference that the agents themselves now act in the coalitions in MCFP-A. We define...

Citation Context

...he discourse/interactions within a group depends on the affective state of a student (Issrof and Soldato, 1996) and his or her social relationship with other students in the group. Jones and Issroff (=-=Issroff and Jones, 2005-=-) and Vass (Vass, 2002) report that, students who are friends have established ways of working which are implicitly understood rather than explicitly discussed. In addition, (Kreijns et al., 2003) men...

Citation Context

...(as perceived by the student) change according to the frequency, extent, and quality of collaboration (e.g., how many times did my peer helped me). 4.1.4 Scaffolding Bruner (Bruner, 1985) and Cazden (=-=Cazden, 1983-=-) define scaffolding as the act of providing assistance to a child so that he or she is able to carry out a task (e.g., solve a problem) that he or she cannot do by herself. Over time, the concept of ...

Citation Context

...n Research I (Research) Solved Problem (Shehory and Kraus, 1995) Coalition formation as a method of task allocation (Sandholm and Lesser, 1997) Coalition formation for computationally bounded agents (=-=Dieckmann and Schwalbe, 1998-=-) Coalition formation in dynamic environments (Sandholm et al., 1999) Computational complexity of coalition formation (Vassileva et al., 2002) Coalition formation for long term coalitions (Breban and ...

Citation Context

...ah and Lesser, 2004) Coalition formation as a method of agent organization (Yokoo et al., 2005) Solution properties of the coalition formation problem in open anonymous environments (e.g., internet) (=-=Bachrach and Rosenschein, 2008-=-) Coalition formation in heterogeneous task solution environment (Sombattheera and Ghose, 2008) Anytime coalition structure generation for coalition formation problems Although these research works di...

Citation Context

... Agent coalition formation under uncertainty. Negotiation-based and Bayesian learning enabled distributed coalition formation protocol was provided. Yes Yes (Wang and Singh, 2007; Erete et al., 2008; =-=Hang et al., 2008-=-, 2009; Hendrix et al., 2009) Modeling the trust of other agents. Various social and communication-strategy-based distributed algorithms are discussed that could capture an agent’s local estimate of t...

Citation Context

...oblem solving in social networking sites (e.g., fund raising for earthquake victims in Haiti through facebook), and effective coalition formation among the users for problem solving in crowdsourcing (=-=Brabham, 2008-=-) Our future work regarding our proposed solution of the MCFP-A problem involves improving the agents’ learning of the environment dynamics with more robust learning algorithms. Such improvements woul...

Citation Context

...ing environments when they cannot solve the task individually. Software system that detects opportunities to form learner groups and a single agent that negotiates learner groups is designed. Yes na (=-=Pechoucek et al., 2002-=-): Forming coalitions for the human users or organizations for solving large, complicated missions. A negotiation-based distributed coalition formation based on the agents’ social knowledge of acquain...

Citation Context

...the simulation run is repeated with Hete-A group formation method. So, the knowledge gain of the students in SimCoL follows patterns described by other CSCL researchers (Roberts and McInnerney, 2007; =-=Clancey, 2004-=-). Figure 4.9: Total expertise gain of students collaborating in groups formed by random group formation method. Correlation with Observed CSCL Results. Here we try to compare our simulation results w...

Citation Context

...and vendors) (Chalkiadakis and Boutilier, 2003) Coalition formation for agent coordination (Dang and Jennings, 2004) Optimization of social payoff in coalition formation 26 (Research) Solved Problem (=-=Abdallah and Lesser, 2004-=-) Coalition formation as a method of agent organization (Yokoo et al., 2005) Solution properties of the coalition formation problem in open anonymous environments (e.g., internet) (Bachrach and Rosens...

Citation Context

...unts. To solve this problem, several researchers (Boongasame et al., 2009; Louta et al., 2008; Chan and Leung, 2008, 2007; Asselin and Chaib-Draa, 2006; Palopoli et al., 2006; Cornforth et al., 2004; =-=Li et al., 2003-=-; Tsvetovat and Sycara, 2000) investigated the buyer coalition formation in the business domain and proposed multiagent frameworks and algorithms where the agents could form coalitions on behalf of th...

Citation Context

... works were not designed to address both of the aspects (ca2, ca3) critical to the solution of MCFP-M. Table 2.1: Agent-Mediated Coalition Formation Research I (Research): Problem. Solution. ca2 ca3 (=-=Ikeda et al., 1997-=-; Inaba et al., 2000): Learner coalition formation in collaborative learning environments when they cannot solve the task individually. Software system that detects opportunities to form learner group...

Citation Context

... how agents learn from each other when they interact or negotiate or collaborate (Chalkiadakis, 2007; Kraus et al., 2003; Chan and Le14 ung, 2008; Tsvetovat and Sycara, 2000; Klusch and Gerber, 2002; =-=Kraus et al., 2004-=-) but not on what learning can do to an agent’s type when forming or joining a coalition, and parlaying that into the future. Our solution, on the other hand, integrally considers that, allowing agent...

Citation Context

...on formation- and scaffolding-related aspects of the agent-mediator cases of the ADLIT. A research approach that investigates the issue of adjustable autonomy could be found in (Screrri et al., 2002; =-=Maheswaran et al., 2004-=-). 35 stochastic–meaning the behavior of the human human users cannot be accurately modeled; only modeled with probability. Assumption 7 (Reward) The rewards achievable by the human coalition can be d...

Citation Context

... by his or her social relationship with other group members. This result portrays a common scenario where students often refuse to collaborate/contribute when they see other members solving the task (=-=Roberts and McInnerney, 2007-=-). This inspires us to include social relationship of the agents in the iHUCOFS framework’s (Section 3.3 in Chapter 3) group formation mechanism. 4.4.2 Compound Impact Analysis This compound impact an...

Citation Context

...e.g., businesses forming coalitions to improve their earnings or to solve difficult tasks ((Khanna and Yafeh, 2007); (Iacobucci and Rosa, 2005)) and political coalition formation to pass legislature (=-=Kreppel and Tsebelis, 1999-=-)) well. 1.2 Solution Approach To solve MCFP-M, we propose the Integrated Human Coalition Formation and Scaffolding Framework (iHUCOFS). iHUCOFS is a multiagent-based coalition formation framework in ...

Citation Context

...buted algorithms are discussed that could capture an agent’s local estimate of the other agents’ performance or behavior for a specific task or tasks in the form of trust. na Yes (Stone et al., 2000; =-=Kuhlmann et al., 2006-=-; Rettinger et al., 2008). Modeling teammates in the robocup soccer. Learning-enabled and statistical modeling of teammates are designed. na Yes na: not addressed Figure 2.2: Categorization of Agent-A...

Citation Context

... for solving large, complicated missions. A negotiation-based distributed coalition formation based on the agents’ social knowledge of acquaintance models was provided. na na (Vassileva et al., 1999; =-=Bull et al., 2001-=-; Greer et al., 2001; Vassileva et al., 2003): Learner coalition formation for providing 1-to-1 peer help in collaborative learning environment. A framework where agents watch and develop probabilisti...

Citation Context

...ifferent non-contributing students in CSCE 475 did not repeat their free-riding behavior in future topic summary experiments. Therefore, these results suggests that the data collected by Researchers (=-=Ebner et al., 2008-=-) 106 mention that the main problem regarding free-riding is two-fold: first, often collaborative learning tools do not allow the teacher to accurately capture student’s contributions toward his or he...

Citation Context

...ng to the principles described in iHUCOFS, we have designed the VALCAM algorithm (Soh, Khandaker, Liu and Jiang, 2006; Soh and Khandaker, 2007; Khandaker et al., 2011; Soh, Khandaker and Jiang, 2006; =-=Khandaker et al., 2006-=-; Soh et al., 2008; Khandaker and Soh, 2011b) and investigated its impact by implementing it in the asynchronous version of I-MINDS (Intelligent Multiagent Infrastructure for Distributed Systems in Ed...

Citation Context

...se on a topic or skill, which could be topic-specific, e.g., learning how to solve differential equation, or topic-independent, e.g., teamwork or communication skills. According to learning theories (=-=Ellis et al., 1997-=-; Shell and Brooks, 2007) the four main elements that affect how a person learns are: (1) what the student already knows (knowledge), (2) how able/intelligent the student is (ability), (3) how motivat...

Citation Context

...t of type θi have participated in a coalition C for task TC , implicit learning action icC , and explicit learning action scC Notice that in Equation 6.15-Equation 6.18, the agents use a frequentist (=-=Amati, 2006-=-) approach where they use their observations to update their estimates of the probabilities. 6.4 Implementation In this section, we first describe how we have adopted the coalition formation problem d...

Citation Context

...walbe, 1998) Coalition formation in dynamic environments (Sandholm et al., 1999) Computational complexity of coalition formation (Vassileva et al., 2002) Coalition formation for long term coalitions (=-=Breban and Vassileva, 2002-=-) Coalition formation in business environments (coalitions between customers and vendors) (Chalkiadakis and Boutilier, 2003) Coalition formation for agent coordination (Dang and Jennings, 2004) Optimi...

Citation Context

...group. Finally, as reported in (Khandaker et al., 2006), the students form their view of other students due to the type and extent of collaboration they receive from their peers. Clear and Kassabova (=-=Clear and Kassabova, 2005-=-) further report that in collaborative learning settings it is common to have students whose motivation is affected by the motivation of other group members. When the other group members are motivated...

Citation Context

...n can be determined by psychometric tests (Vallerand et al., 1992). However, these psychometric tests themselves often suffer from measurement errors which reduces the accuracy of their measurements (=-=Carmines and Zeller, 1979-=-). 3.2 Design Principles According to our assumptions of the iHUCOFS framework (Section 3.1) we use the following principles while designing the MHCF algorithm for coalition formation: 36 User Modelin...

Citation Context

...y obtaining volume discounts. To solve this problem, several researchers (Boongasame et al., 2009; Louta et al., 2008; Chan and Leung, 2008, 2007; Asselin and Chaib-Draa, 2006; Palopoli et al., 2006; =-=Cornforth et al., 2004-=-; Li et al., 2003; Tsvetovat and Sycara, 2000) investigated the buyer coalition formation in the business domain and proposed multiagent frameworks and algorithms where the agents could form coalition...

Citation Context

...ier, 2008): Repeated Agent coalition formation under uncertainty. Negotiation-based and Bayesian learning enabled distributed coalition formation protocol was provided. Yes Yes (Wang and Singh, 2007; =-=Erete et al., 2008-=-; Hang et al., 2008, 2009; Hendrix et al., 2009) Modeling the trust of other agents. Various social and communication-strategy-based distributed algorithms are discussed that could capture an agent’s ...

Citation Context

...rds over time. To solve MCFP-M problem, we have proposed the iHUCOFS framework and to solve MCFP-A, we have proposed the ADLIT framework for 184 coalition formation. iHUCOFS (Khandaker and Soh, 2008; =-=Soh and Khandaker, 2007-=-) is a multiagent-based coalition formation framework in which a set of intelligent agents assist a set of human users form coalitions and scaffold those human users to optimize their effectiveness an...

Citation Context

...ences into account in their framework or solution approach. For example, the agents in the solutions proposed in (Kraus et al., 2003); (Chalkiadakis and Boutilier, 2004); (Hosam and Khaldoun, 2006); (=-=Chalkiadakis, 2007-=-) do not consider the impact of an agent influencing type changes in its coalition members by learning actions. Due to the necessary collaboration of the agents in the coalition formation problems, if...

Citation Context

... the students through learning so that, they are able to earn higher rewards for future tasks. To solve MCFP-A, our proposed ADLIT framework allows an agent to make a series of sequentially rational (=-=Chalkiadakis and Boutilier, 2008-=-) decisions to improve the tradeoff between the current task and future task to maximize its rewards over time with respect to its type. In our design and implementation of ADLIT, we formulate the typ...

Citation Context

...dinator of the CSCL sessions. The teacher delivers instructions, forms groups, and assigns collaborative tasks. In SimCoL, we have implemented three different group formation methods: random, Hete-A (=-=Gogoulou et al., 2007-=-), and VALCAM (Soh, Khandaker and Jiang, 2006) group formation method. Algorithm 4.1 shows how the teacher carries out the CSCL session through a set of simulation steps. First, the teacher initialize...

Citation Context

...rtant but yet to be automatically solved real-world coalition formation prob8lems e.g., businesses forming coalitions to improve their earnings or to solve difficult tasks ((Khanna and Yafeh, 2007); (=-=Iacobucci and Rosa, 2005-=-)) and political coalition formation to pass legislature (Kreppel and Tsebelis, 1999)) well. 1.2 Solution Approach To solve MCFP-M, we propose the Integrated Human Coalition Formation and Scaffolding ...

Citation Context

...d information for that student to help him or her solve that task. Researchers in the CSCL community are now utilizing scaffolding in the form of incorporating structure of learning activities (e.g.,(=-=Harrer et al., 2008-=-)) and improving peer support (e.g.(Liu et al., 2008)). As CSCL researchers (e.g., (Khandaker et al., 2006; Liu et al., 2008) note that due to being in different zones of proximal development, the lea...

Citation Context

... where agents watch and develop probabilistic models of the users and communicate among themselves to find the best 2-member coalition was developed. na Yes 21 (Research): Problem. Solution. ca2 ca3 (=-=Li et al., 2006-=-): Forming coalition of learners with similar expertise and preferences. A fuzzy set theory based framework forms coalitions based on the agents’ model of the learners is developed. Yes na na: Not add...

Citation Context

... that task. Researchers in the CSCL community are now utilizing scaffolding in the form of incorporating structure of learning activities (e.g.,(Harrer et al., 2008)) and improving peer support (e.g.(=-=Liu et al., 2008-=-)). As CSCL researchers (e.g., (Khandaker et al., 2006; Liu et al., 2008) note that due to being in different zones of proximal development, the learners benefit most when the scaffolding is targeted ...

Citation Context

... heterogeneous objectives in a dynamic, uncertain, and noisy environment. Clan(trusted agent group) based coalition formation mechanism was proposed. na Yes 28 (Research): Problem. Solution. ca1 ca2 (=-=Blankenburg and Klusch, 2004-=-): Forming coalition of the most reliable agents in repeated coalitional games. A trusted kernel-based, encryptionenabled coalition formation protocol was designed. Yes na (Gaston and DesJardins, 2005...

Citation Context

...gned to address both of the aspects (ca2, ca3) critical to the solution of MCFP-M. Table 2.1: Agent-Mediated Coalition Formation Research I (Research): Problem. Solution. ca2 ca3 (Ikeda et al., 1997; =-=Inaba et al., 2000-=-): Learner coalition formation in collaborative learning environments when they cannot solve the task individually. Software system that detects opportunities to form learner groups and a single agent...

Citation Context

... scaffolding to the students using agents is a well-researched area in Intelligent Tutoring systems. We plan to incorporate the existing intelligent tutoring systems research (Jung and VanLehn, 2010; =-=Arroyo et al., 2010-=-) in the mediator agents’ reasoning so that they are able to explicitly scaffold the human users. Furthermore, to improve the agent learning capability in the MHCF algorithm, we plan to study the impa...

Citation Context

...n under uncertainty. Negotiation-based and Bayesian learning enabled distributed coalition formation protocol was provided. Yes Yes (Wang and Singh, 2007; Erete et al., 2008; Hang et al., 2008, 2009; =-=Hendrix et al., 2009-=-) Modeling the trust of other agents. Various social and communication-strategy-based distributed algorithms are discussed that could capture an agent’s local estimate of the other agents’ performance...

Citation Context

... of goods and have limited bargaining power or (2) are interested in increasing their profit margin by obtaining volume discounts. To solve this problem, several researchers (Boongasame et al., 2009; =-=Louta et al., 2008-=-; Chan and Leung, 2008, 2007; Asselin and Chaib-Draa, 2006; Palopoli et al., 2006; Cornforth et al., 2004; Li et al., 2003; Tsvetovat and Sycara, 2000) investigated the buyer coalition formation in th...

Citation Context

...2) are interested in increasing their profit margin by obtaining volume discounts. To solve this problem, several researchers (Boongasame et al., 2009; Louta et al., 2008; Chan and Leung, 2008, 2007; =-=Asselin and Chaib-Draa, 2006-=-; Palopoli et al., 2006; Cornforth et al., 2004; Li et al., 2003; Tsvetovat and Sycara, 2000) investigated the buyer coalition formation in the business domain and proposed multiagent frameworks and a...

Citation Context

... need to buy small amount of goods and have limited bargaining power or (2) are interested in increasing their profit margin by obtaining volume discounts. To solve this problem, several researchers (=-=Boongasame et al., 2009-=-; Louta et al., 2008; Chan and Leung, 2008, 2007; Asselin and Chaib-Draa, 2006; Palopoli et al., 2006; Cornforth et al., 2004; Li et al., 2003; Tsvetovat and Sycara, 2000) investigated the buyer coali...

Citation Context

...e. Providing of implicit scaffolding to the students using agents is a well-researched area in Intelligent Tutoring systems. We plan to incorporate the existing intelligent tutoring systems research (=-=Jung and VanLehn, 2010-=-; Arroyo et al., 2010) in the mediator agents’ reasoning so that they are able to explicitly scaffold the human users. Furthermore, to improve the agent learning capability in the MHCF algorithm, we p...

Citation Context

..., and Assumption 7 (Section 3.1). 3.3 MHCF Algorithm Based on the design principles (Section 3.2) of the iHUCOFS framework, we have designed the Multiagent Human Coalition Formation (MHCF) algorithm (=-=Khandaker, 2009-=-; Khandaker and Soh, 2010b,c) for forming human coalitions. 3.3.1 Environment The MHCF environment E is denoted as a 5-tuple 〈S,A,G, T,R〉. Here, 38 • S = {s1, . . . , sns} is the set of human users • ...

Citation Context

...imited bargaining power or (2) are interested in increasing their profit margin by obtaining volume discounts. To solve this problem, several researchers (Boongasame et al., 2009; Louta et al., 2008; =-=Chan and Leung, 2008-=-, 2007; Asselin and Chaib-Draa, 2006; Palopoli et al., 2006; Cornforth et al., 2004; Li et al., 2003; Tsvetovat and Sycara, 2000) investigated the buyer coalition formation in the business domain and ...

Citation Context

...ques of such problems. Due to their ability to communicate, cooperate, and make autonomous decisions, multi-agent systems are gaining popularity as a tool for designing real-world software solutions (=-=Fischer et al., 2007-=-). Many of those real-world problems e.g., task allocation, bargain buying in business transactions, resource allocation, and human coalition formation require formation of coalitions in uncertain env...

Citation Context

... of effectiveness (in terms of improving the types of the agents) of the implicit and explicit learning interactions performed by the agents. Our assumption is driven by the learning theory research (=-=Hoppe et al., 2003-=-) that describe the variances among the impact of different learning actions (learning by being taught, learning by apprenticeship) and the relative difference in the teacher’s and learner’s expertise...

Citation Context

... agents’ type-changing influences into account in their framework or solution approach. For example, the agents in the solutions proposed in (Kraus et al., 2003); (Chalkiadakis and Boutilier, 2004); (=-=Hosam and Khaldoun, 2006-=-); (Chalkiadakis, 2007) do not consider the impact of an agent influencing type changes in its coalition members by learning actions. Due to the necessary collaboration of the agents in the coalition ...

Citation Context

...ign or implementation in I-MINDS in this dissertation. For a summary of our achievements regarding VALCAM and its implementation in the asynchronous version of I-MINDS, see (Khandaker and Soh, 2011b; =-=Khandaker et al., 2011-=-). 9To understand the validity and implications of iHUCOFS framework, we have designed SimCoL (Khandaker and Soh, 2010d) - a simulation environment in which a set of agents, guided by the published co...

Citation Context

...ning actions (Equation 6.27), and (3) the learning cost (Equation 6.26). • Scalability — Researchers define scalability of multiagent systems as the overall effect of processing times on performance (=-=Rana and Stout, 2000-=-) when the size and scale of the system increases. In this subsection, we aim to establish that, the performance improvement (in terms of the utility) that can be gained by using ADLIT’s coalition for...

Citation Context

..., 2010d)—a multiagent tool for simulating student learning interactions in a CSCL environments. SimCoL uses a set of observations and theories published in the collaborative learning research domain (=-=Shell and Brooks, 2007-=-) to simulate the student-student and student-teacher interactions in a CSCL classroom. The analysis of the results of our conducted simulation experiments (Khandaker and Soh, 2010d) shows that SimCoL...