Top PDF Agent Based Models of Competition and Collaboration

Agent Based Models of Competition and Collaboration

Agent Based Models of Competition and Collaboration

3.1 Motivation Distributed systems are an active and important eld with applications ranging from distributed heterogeneous computing systems [32] to mobile sensor networks [45], and their performance is dependent on the coordination of disparate sub-systems to full the overall goal of the system. This problem can be seen as one of eciently deploying a nite set of resources in order to complete a distributed set of sub-tasks, where these sub-tasks further this overall goal. It is clear that, in theory, the best method for coordinating these resources must be centralised. A central controller can, at minimum, issue commands causing resources to be deployed as if according to the optimal decentralised behaviour. In fact, access to global information and the ability to coordinate agents should allow better performance than any collection of individuals. However, limitations on resources such as computational power and/or communication costs [36] mean that centralised solutions are not ecient in practice. This is particularly true for large systems as the calculation time of an optimal allocation of tasks becomes a major limitation [13]. Large systems also decrease the eectiveness of global inter-agent communication. Shehory et al. [65] point out that if n agents are communicating with each other, this involves a total of O(n 2 )
Show more

141 Read more

Institutional collaboration and competition in community-based education

Institutional collaboration and competition in community-based education

Summary We sought to determine whether competition for community-based training sites exists among health professions schools, and to examine faculty and senior administrators’ perspectives on institutional collaboration for community-based education. Eight academic health centers (AHCs) in the USA were selected by objective criteria for their significant community involvement. Chief executive officers, vice chancellors, deans, and the individuals responsible for community-based education, research and community service responded to written surveys. The overall response rate was 79% (n = 91). Responses were subjected to quantitative and qualitative analyses. Leaders of community- based education reported that ‘competition for community-based training sites’ is a barrier to community involvement. ‘Competition for community-based training sites’ was positively related to ‘call for increasing percentage of graduates to enter primary care careers’ (0.30, p 5 0.01) and negative related to ‘collaboration exists between the community and your school/AHC’ ( 7 0.28, p 5 0.05). Respondents reported that a moderate level of collaboration across schools exists. While medical school respondents reported having collaborative relationships with other health professions schools and with the community, nursing respondents reported medicine’s performance at a significantly lower level. Public health and nursing faculty reported that they are competing with medical schools for sites they had traditionally used for their students. Competition for sites is an unintended outcome of the increased emphasis on community-based education in health professions curricula. We recommend AHCs form joint committees across schools to effectively address community-based sites as a limited resource, and to consider a wider range of community-based organizations as training partners.
Show more

12 Read more

Agent Based Models of Competition and Collaboration

Agent Based Models of Competition and Collaboration

3.1 Motivation Distributed systems are an active and important eld with applications ranging from distributed heterogeneous computing systems [32] to mobile sensor networks [45], and their performance is dependent on the coordination of disparate sub-systems to full the overall goal of the system. This problem can be seen as one of eciently deploying a nite set of resources in order to complete a distributed set of sub-tasks, where these sub-tasks further this overall goal. It is clear that, in theory, the best method for coordinating these resources must be centralised. A central controller can, at minimum, issue commands causing resources to be deployed as if according to the optimal decentralised behaviour. In fact, access to global information and the ability to coordinate agents should allow better performance than any collection of individuals. However, limitations on resources such as computational power and/or communication costs [36] mean that centralised solutions are not ecient in practice. This is particularly true for large systems as the calculation time of an optimal allocation of tasks becomes a major limitation [13]. Large systems also decrease the eectiveness of global inter-agent communication. Shehory et al. [65] point out that if n agents are communicating with each other, this involves a total of O(n 2 )
Show more

141 Read more

Remodelling the third sector: advancing collaboration or competition in community based initiatives?

Remodelling the third sector: advancing collaboration or competition in community based initiatives?

A further consequence of the move to commissioning children and young people’s services has been the introduction of a common contract, containing performance indicators derived from central government targets, which many providers regarded as peripheral to their core work. Recently the government has stated its intention to review top–down imposed targets to allow more local dis- cretion, but ministers are quoted (Carvel, 2007 ) as advocating broader shared tar- gets across services. This study found that community providers felt increasingly vulnerable faced with a shift towards more generic performance targets, since commonality reduced the visibility and demonstrable value of their work and achievements, potentially disadvantaging them in future competition for funds. The effects of performance cultures in shaping the nature of services have been documented elsewhere (Power, 1999 ; Clarke et al., 2000 ) but the community context highlights a particular contradiction between promoting creative local solutions to social problems, and defining the outcomes required in competitively allocated contracts. Where success is judged through measures determined by funders, the need to compete for future contracts induces community providers to conceal failures and accentuate positive outcomes. This concealment, however, serves the interests of mainstream agencies in demonstrating successful policy outcomes. Moreover, if outcomes appear positive, there is less reason to challenge the planning, funding and performance frameworks within which the work is conceived, generating a conspiracy of denial.
Show more

23 Read more

Evolutionary mechanism design using agent-based models

Evolutionary mechanism design using agent-based models

in Duffy and Unver (2008) are ascending-bid format, second-price format, and a hybrid of the ascending-bid English auction and the second-price sealed bid auction. Different from the traditional auctions, online auctions allow bidders to bid anywhere in the world at any time with Internet. Some online auctions run many days, providing a good chance for more bidders to notice and enter the auction. Lucking-Reiley et al. (1999) and Hasker et al. (2004) express that the longer online auction runs, the more bidders are attracted with higher prices. As found by Lucking-Reiley et al. (1999), prices are 24% higher, on average, in 7-day online auctions than in short period auctions. However, auctions with longer duration raise the question of the optimal timing of the bid. An interesting phenomenon is that of “late” bidding, which is placed very close to the end of the auction, gives other bidders little time to respond. Another design question is whether bids in online auctions should be sealed or open, i.e., whether a bidder has a chance to observe how others’ bidding activities evolve, and reacts to their bidding strategies during the auction process. Ivanova-Stenzel and Salmon (2004) show that in a private-value environment open bidding is much preferred by laboratory subjects. In ad- dition, Ariely et al. (2005) conduct a laboratory human subject experiment which shows that the speed of learning the optimal bidding strategy is much faster in open auctions than in sealed-bid auctions. However, there are some other argument against the superior of open format. Cramton (1998) re- ports that open bidding fails to generate high revenues when bidders are risk-averse and competition is low. Moreover, Klemperer (2004) finds that collusion among bidders in open auctions may reduce the revenue.
Show more

274 Read more

An agent-based model of supply chain collaboration: Investigating manufacturer loyalty

An agent-based model of supply chain collaboration: Investigating manufacturer loyalty

The experiments performed in this paper consider only one aspect of supply chain collaboration in the market, supply chain loyalty. As a next step we plan to consider further aspects of collaboration and competition behaviour, including the distance of manufacturer strategic movement, supplier loyalty, and the supplier’s maximum number of relationships. The results are also analysed by disregarding the interaction between collaboration factors, such as duration of collaboration and number of sourcing. A similar analysis is also under way considering supply chain competition. Varying the model content, such as supply and demand characteristics, will also be simulated in the future in order to gain a better understanding of collaboration in the supply chain. Moreover, several case studies will be used to test the model.
Show more

10 Read more

Deep Learning in Agent-Based Models: A Prospectus

Deep Learning in Agent-Based Models: A Prospectus

Applications of ANNs to time series forecasting problems in economics include: financial market forecasting (Trippi, 1993; Azoff, 1994; Refenes, 1995; Gately, 1996), foreign exchange rates (Weigend, 1992; Refenes, 1993; Kuan and Liu, 1995), load demand forecasts on electricity markets (Bacha and Meyer, 1992; Srinivasan et al., 1994), commodity prices (Kohzadi et al., 1996), and macroeconomic indices (Maasoumi, 1994) A review of applications of ANNs in the field of Management Science and Operations Research is given by Wilson (1992) and Sharda (1994). The M-competition (Makridakis et al., 1982) provides a widely cited data base for comparing the forecasting performance of ANNs in comparison to traditional statistical methods. The data for the M-competition are mostly from business, economics, and finance, see Kang (1991); Sharda (1994); Tang and Fishwick (1993) for examples. Another comparison is provided by the Santa Fe forecasting competition (Weigend and Gershenfeld, 1993) which includes very long time series coming from various fields.
Show more

19 Read more

PROCESS MODELS FOR AGENT-BASED DEVELOPMENT

PROCESS MODELS FOR AGENT-BASED DEVELOPMENT

A number of specific models may be accommodated under the evolutionary approach (e.g., incremental implementation model, incremental development and delivery model, evolutionary prototype model, etc.). Probably, the most recent model in this direction, and perhaps one of the most interesting, is the Extreme Programming (XP) process model (Beck, 1999) that is being increasingly used in projects with uncertain or changing requirements. XP is an example of “agile approach” (see the agile manifest at www.agilemanifesto.org ) aimed at supporting changes and rapid feedbacks. Agile approaches have become enough popular in the last years, their philosophy can be resumed by their fundamental strategies reported in the agile manifesto, which prescribed to give consideration to: (i) individuals and interactions over processes and tools; (ii) working software over comprehensive documentation; (iii) customer collaboration over contract negotiation; (iv) responding to change over following a plan.
Show more

30 Read more

Agent-Based Model of Price Competition and Product Differentiation on Congested Networks

Agent-Based Model of Price Competition and Product Differentiation on Congested Networks

There are several main assumptions in the proposed model. First of all, residential locations of all users are assumed to be fixed for simplicity and there are no land use changes. This may result in prediction biases, espe- cially when the model is applied to produce long-term estimates due to the well-known land use–transportation interactions. This fixed land use, and therefore fixed overall travel demand, allows the model to achieve a long-term system equilibrium more easily. However, future research on long-term transportation network dynamics should strive to relax this assumption and use integrated land use–transportation models. Over a long period of time, the impact of transportation investments on land use could be significant. Any land-use changes will in turn influence subsequent transportation network dynamics. Each traveller in the model is assumed to have a predetermined travel budget, which is a simplified representation of the underlying trade-off between reducing travel cost and reaching premium activity locations. Two types of learning are considered in the model: learning from doing, and learning from information exchange. All information exchanges in the model occur exclusively and locally at nodes in the transportation systems, which is structurally appealing for model development but does not directly capture global information sharing (for example, through mass media). These and other minor model assumptions are further discussed and justified in the following sections.
Show more

27 Read more

Agent based services for the validation and calibration of multi agent models

Agent based services for the validation and calibration of multi agent models

handling complexity (Jennings, 2000). The overall problem to be tackled can be decomposed into a number of autonomous agents each with its own capabilities and goal directed behaviour. The designed collaboration between the agents is the means by which the relationships between the decomposed elements of the overall problem can be managed. Thus agent-based services can themselves be engineered as multi-agent tools and thus, if necessary, developed and enhanced incrementally. The network mobility and interoperability of agents opens the possibility for agent-based services to be made available on the Internet as distributed components. This is the approach that underscores our investigation of agent-based services directed towards the validation of multi- agent models. The efficacy of such a solution has already been demonstrated in an investigation of network-resident agent-based services to provide data quality analyses for numerical simulation models that are loosely coupled with GIS (Li Y., 2006). As presented in the following case study, two collaborating agents have been designed as agent-based services for validating multi-agent models: one to carry out sensitivity analyses, the other to assist in model calibration. Although as discussed in the previous Section, a range of issues need to be addressed in establishing the validity of multi-agent models, the function of the agent-based services presented here, whilst not encompassing all conceivable aspects of model validation, nevertheless represent two key elements and are thus sufficient for a proof-of-concept.
Show more

23 Read more

Analysis of collusion and competition in electricity markets using an agent-based approach

Analysis of collusion and competition in electricity markets using an agent-based approach

Liu and Hobbs [69] introduce a framework for modeling tacit collusion in which Gen- Cos collectively maximize a Nash bargaining objective function in their study on a com- petitive pool-based electricity market operated by the ISO. To the best of our knowledge, this is the first study that considers network congestion in modeling tacit collusion; they propose MPEC and EPEC models. Although EPEC models are versatile, they are widely known for two issues. First, an equilibrium may not exist; second, it is hard to com- pute. Therefore, some heuristic algorithms are proposed to solve these models. Besides, the numerical solution of EPECs is a novel area with only few numerical studies [65]. We present an alternative formulation which is relatively easier to work with and can be handled with linear programming through some assumptions.
Show more

132 Read more

VidyutVanika: A Reinforcement Learning Based Broker Agent for a Power Trading Competition

VidyutVanika: A Reinforcement Learning Based Broker Agent for a Power Trading Competition

Since 2012, several research groups have benchmarked, de- ployed and published strategies using Power TAC. ¨ Ozdemir and Unland (2015; 2018a; 2018b), Power TAC 2014 & 2017 Winners, use Genetic Algorithm and aggressive pric- ing to design tariffs for the tariff market, while using adaptive Q-learning in the wholesale market. They also predict the demand of customers using a combination of SARIMA and AR models. Power TAC 2015 Winners, Ur- ban and Conen, design their TOU Tariff rates using a Hill Climbing algorithm. Past Power TAC participants R´ubio et al. (2015) present a fuzzy-logic based trading mechanism, while Liefers, Hoogland, and La Poutr´e (2014) use a heuris- tic inspired from Tit-For-Tat strategy in Iterated Prisoner’s Dilemma, to determine tariff rates based on competing tar- iffs. Chowdhury et al. (2017, 2018) use an MDP & Q- Learning based tariff market strategy and a Monte-Carlo Tree Search based wholesale strategy, with the former in- corporating the market share and cash position of the agent into the state space while taking actions on maintaining, in- crementing or decrementing tariff rates.
Show more

8 Read more

A Platform for Trading Agent Competition

A Platform for Trading Agent Competition

Trading Agent Competition Market Design Game Strategic Trader Game Conclusion and Future Work Conclusion and Future Work Summary Automated trading has been a rapidly growing research area. Our system allows more variety and sophisticated trading strategies to be created and tested.

13 Read more

The 2001 Trading Agent Competition

The 2001 Trading Agent Competition

ATTac uses a predictive, data-driven approach to bid based on expected marginal values of all available goods. A price-predictor based on boosting techniques [11, 12] is at the heart of the al- gorithm. This price-predictor generates distributions over expected hotel closing prices. ATTac then samples from these distributions in an effort to compute the expected marginal utility of each good. It then bids exactly these expected marginal utilities. As the game proceeds, the price distri- butions change in response to the observed price trajectories, thus causing the agent to continually revise its bids. Note that by using this strategy, provided that the price is right, ATTac automatically buys contingency goods to guard against the possibility of the most desired goods becoming too expensive.
Show more

15 Read more

An agent-based Web service workflow model for inter-enterprise collaboration

An agent-based Web service workflow model for inter-enterprise collaboration

Since the current implementation only deals with the partner search and selection, the Contract Net Protocol or its modified version is enough for the coordination among the broker agent and supplier agents. Other Web services standards such as BPEL4WS will be adopted for further interactions/collaborations between business part- ners. Moreover, software agents for intra-enterprise sys- tems are simulated with the support of the Autonomous Agent Development Environment (AADE). AADE (Hao, Shen, & Zhang, 2005) is a FIPA (FIPA specifications) com- pliant engineering-oriented agent framework developed at the National Research Council Canada’s Integrated Man- ufacturing Technologies Institute and it has been used for the implementation of an agent-based dynamic manufac- turing scheduling system at the enterprise level, where each machining tool is represented by one resource agent (machine agent). Details of the agent-based manufacturing resource scheduling system implemented using AADE plat- form are described in Wang, Shen, and Hao (2005).
Show more

13 Read more

Duration of collaboration from a market perspective: An agent-based modeling approach

Duration of collaboration from a market perspective: An agent-based modeling approach

In addition, several studies suggest that long-term relationships lead the partners to be more likely dissatisfied with the cooperative arrangement. Several studies in marketing management found empirically that long term relationships can lower trust and the service performance (Grayson and Ambler 1999). This finding is supported by studies in strategic management which posits that close partnerships can encourage the collaborating partner to be too dependent on each other (Inkpen and Beamish 1997). In addition, when a firm has a better understanding of what the other knows through a high degree of information sharing between parties, the partnerships become unstable and fragile. This risk will be more apparent if the demand uncertainty is very high or very low (Sun and Debo 2014). In the case of Japanese car manufacturers, Dyer and Ouchi (1993) find that a long-term collaboration does not necessarily need a very high involvement of collaborating firms. In addition, the long-term relationships strategy is also often doubted by SCM practitioners. This strategy is viewed to be risky as developing and maintaining trust between firms are difficult in business relationships. This negative opinions often come from suppliers with larger scale buyers who aggressively established strategic and long-term cooperation with them (Bensaou 1999).
Show more

11 Read more

Agent-based models and individualism: is the world agent-based?

Agent-based models and individualism: is the world agent-based?

of a range of views on the increased fragmentation and individualism of social life. Whichever view is most compelling, there does appear to be general agreement over a need to account for the increasing importance accorded individual preference, opinion, and choice. In the current context, the important point is that the intellectual current is flowing strongly in favour of individualist models. We have seen, in this section, that it is possible to relate the proliferation in the use of agent-based models to two sources: advances in computer technologies which enable the creation of such models, and an intellectual climate ready to sustain individual-based models and open to claims that we should expect the emergence of social structures from such models. The inter-influence between an individualist view of society as (merely) the aggregate behaviour of large groups of individuals, and software agents and DAI — and through these to agent-based modelling, is clear in the collection The Ecology of Computation (Huberman, 1988). Here, the ideas of Dawkins, Hayek and the ecological and evolutionary views of markets are connected and formalised so that they can be adequately modelled using computers.
Show more

20 Read more

Markov chain aggregation for agent-based models

Markov chain aggregation for agent-based models

Time-series-based aggregation schemes as the one proposed by Görnerup and Jacobi (2008) as well as ǫ-machines (Crutchfield and Young, 1989; Shal- izi and Crutchfield, 2001) are appealing from the theoretical point of view, but their application to ABM aggregation is limited by their computational complexity (cf. Görnerup and Jacobi, 2008, 13). The fact that, even in very simple ABMs, the state space of the process to be handled becomes very large, challenges these approaches in two ways. The first one concerns the »combinatorial explosion« (ibid., 11) of the number of possible partitions, which is in fact a general difficulty in lumpability whenever the partition is not given a priori (see Sec. 2.4.1). More importantly, however, the larger the alphabet (and Σ becomes really large!), the more data must be generated and evaluated in order to obtain a workable approximation of the probabil- ity distribution of sequence blocks (cf. Shalizi and Crutchfield, 2001, Sec. VII.B/C). One way around this problem is to restrict to block size to one, as in Shalizi et al. (2004), which is actually exact if the original process is a Markov chain. Still, in this case, the number of states is huge and the estimation of the conditional probabilities (on the basis of which equivalence classes are constructed) requires a lot of simulation data.
Show more

165 Read more

Enhancing Agent-Based Models with Discrete Choice Experiments

Enhancing Agent-Based Models with Discrete Choice Experiments

Abstract: Agent-based modeling is a promising method to investigate market dynamics, as it allows modeling the behavior of all market participants individually. Integrating empirical data in the agents’ decision model can improve the validity of agent-based models (ABMs). We present an approach of using discrete choice experi- ments (DCEs) to enhance the empirical foundation of ABMs. The DCE method is based on random utility theory and therefore has the potential to enhance the ABM approach with a well-established economic theory. Our combined approach is applied to a case study of a roundwood market in Switzerland. We conducted DCEs with roundwood suppliers to quantitatively characterize the agents’ decision model. We evaluate our approach us- ing a fitness measure and compare two DCE evaluation methods, latent class analysis and hierarchical Bayes. Additionally, we analyze the influence of the error term of the utility function on the simulation results and present a way to estimate its probability distribution.
Show more

22 Read more

Optimization and Control of Agent-Based Models in Biology: A Perspective.

Optimization and Control of Agent-Based Models in Biology: A Perspective.

A well-known family of ABMs known as Sugarscape (Wilensky 2009; Epstein and Axtell 1996) has been used for the study of a variety of control processes in the life sci- ences, social science, and economics. The stochastic Sugarscape ABMs include agent heterogeneity, environmental heterogeneity, and accumulation of agent resources (i.e., sugar) over time, thus incorporating the main complexities frequently found in ABMs. Agents negotiate a spatial environment in search of a resource called sugar, with higher sugar concentrations represented as elevations in the landscape. Different agents have differing abilities to perceive sugar gradients, leading to different levels of agent fit- ness. Complete lack of sugar leads to agent death. Control is included as taxation of agents’ sugar resources, with the goal of maximizing a weighted combination of total taxes less a measure of the impact of taxation on the population. Recently, Christley et al. (2015) approximated a Sugarscape model using a system of parabolic PDEs. The goal was to explore optimal control scenarios for Sugarscape, applying mathematical optimization approaches to the PDE model. This approach performed well in scenar- ios in which the control was assumed to be constant. Optimal controls generated by applying optimal control theory to the PDE system provided time-varying tax rates specific to an agent’s location and current wealth. When implemented in the ABM, the optimal controls performed reasonably well even though some error was introduced between the PDE and ABM systems.
Show more

26 Read more

Show all 10000 documents...