I. A. Ferguson. TouringMachines: An Architecture for Dynamic, Rational, Mobile Agents. Phd thesis, University of Cambridge, 1992.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....control and monitor an agent s own behaviors are reused to monitor, interpret, and predict its partner s behaviors. Ferguson s work on the TouringMachine architecture focused on agents in dynamic, multi agent environments also involves an explicit modeling layer for tracking other agents[9]. A second area of related work is research specifically focused on agent modeling and plan recognition. Section 1 has discussed some of this work. In addition, some formal approaches for agent modeling, and in particular for recursive agent modeling, are also being investigated[11] Vidal and ....

I. A. Ferguson. TouringMachines: An architecture for dynamic, rational, mobile agents. PhD thesis, University of Cambridge, 1992.

....of action. Non deliberative architectures operate in a stimulus response fashion. One major characteristic of these architectures is the lack of knowledge representations. In contrast to deliberative agents, which are designed to produce optimal actions, reactive agents produce robust actions [17]. Hybrid agent architectures combine the good qualities of both deliberative and reactive architectures i.e. fast response and the generation of optimal course of action. 6 Proposed Agent Architecture The agent architecture presented below is believed to provide a mechanism to implement a ....

Ferguson, I.A.: TouringMachines: An Architecture for Dynamic, Rational, Mobile Agents. PhD Thesis, University of Cambridge, UK (1992)

....the Action Executor and a number of reactive entities called behaviours by robot researchers. The latter ones implement reactive control, whereas the Planner and the Plan Executor support (symbolic) reasoning. This particular architecture has been in uenced largely by work presented in [1, 5] and [2]. 2.1 The World Model The World Model represents objects extracted from physical environment and relations between them. Every object includes at least its symbolic representation. In addition other types of data, like measure 3 ments and locations of physical objects, may be included, if it is ....

I. A. Ferguson. TouringMachines: An Architecture for Dynamic, Rational, Mobile Agents. PhD thesis, Technical Report TR273, University of Cambridge, Computer Laboratory, Cambridge CB2 3QG, England, November 1992.

....for interacting (social) agents [12] A considerable amount of research effort has been dedicated to the exploration of how combination of these architectures can be reconciled in so called hybrid or layered architectures. Layering can be horizontal as in the Touring Machine agent architecture [7, 8] or vertical as in the INTERRAP architecture [12] Under a stronger notion of agency, the BDI theory [1] requires the internal processing state of an agent to be represented by means of a set of a set of beliefs, desires, and intentions. Several agent architectures are based on this theory, ....

I. Ferguson. Touring Machines: An Architecture for Dynamic, Rational, Mobile Agents. Ph.D. Thesis, Computer Laboratory, University of Cambridge, UK. 1992.

....(see, for example: 1] 6] 5] 3] we are particularly concerned with layered ones. In horizontally layered architectures each software layer is directly connected to the sensory inputs and action outputs. A great advantage of horizontally layered architectures (e.g. TOURINGMACHINES [4]) is their conceptual simplicity. However, as the layers are competing with one another, there is the possibility that the overall behavior of the agent will become incoherent. In order to ensure consistency, such kind of architectures typically includes a mediator function, devoted to select the ....

Ferguson, I. A., TouringMachines: An Architecture for Dynamic, Rational, Mobile Agents. PhD thesis. Clare Hall, University of Cambridge, UK (1992)

....Although true autonomy is still a very open research subject, it is advantageous to any agent that has to deal with unexpected situations. Arguments in favour of true autonomy have been put forward in several diverse fields, for example robotics, animal robotics 1 (McFarland 1994) agents theory (Ferguson 1992) and interactive virtual environments (Blumberg 1995) In robotics, emotions are often used to modulate activity in a fixed controller (Ca namero 1997; Bates, Loyall, Reilly 1992a) The social role of emotions has been particularly explored. The external demonstration 1 Modelling of animal ....

Ferguson, I. A. 1992. Touring Machines: An architecture for dynamic, rational, mobile agents. Ph.D. Dissertation, University of Cambridge.

....is some sort of planner. It is worth noting that most other existing robot architectures have two or three layers (PA layer and planner or PA layer, sequencer and planner) There has been no compelling case given in the literature for more layers. Some architectures have used more layers, 71] 3][26] for example, but these have either never been implemented for any real world task that makes use of these extra layers [71] 26] divide the PA layer tasks into extra layers [3] or are not designed to control mobile robots [3] 1.3.3 Agent Perception Representation of a dynamic environment ....

....planner or PA layer, sequencer and planner) There has been no compelling case given in the literature for more layers. Some architectures have used more layers, 71] 3] 26] for example, but these have either never been implemented for any real world task that makes use of these extra layers [71][26], divide the PA layer tasks into extra layers [3] or are not designed to control mobile robots [3] 1.3.3 Agent Perception Representation of a dynamic environment necessarily means updating the representation as the world changes. This makes perception an important design consideration for any ....

Ferguson, I.A. 1992. Touring Machines: An Architecture for Dynamic, Rational, Mobile Agents. Ph.D. Dissertation. Computer Science Department. University of Cambridge.

....in co operating with other actors. 4. 2 Global architecture The architecture (Figure 2) we need contains four modules (Perception, Action, Control and Cognition) distributed in three layers (Reactive, Control and Cognitive) The reactive layer is similar (but extended) to the Touring Machines [7], consisting in perception and action subsystems that interface with the environment of the agent and three control layers (reactive, planning and modelling layers) Similarly, our architecture contains also a reactive layer, with direct association between perception and action modules, and a ....

Ferguson, I. A. (1992) TouringMachines: An Architecture for Dynamic, Rational, Mobile Agents. PhD Thesis, Clare Hall, University of Cambridge.

....that it groups similar types of computations into layers (shown in Figure 1) as opposed to the much more common approach where a functional decomposition into layers is typical. Literally dozens of layered architectures have been recently proposed for autonomous system implementation (see e.g. [1, 3, 5, 8, 13] for just a few of them; good overviews can be found in [2, 38, 22] Although there are di erences both in the way of assigning various tasks to di erent layers and in the way the overall control of the system is executed, the general conclusion is that such layering is bene cial, if not ....

....is virtually impossible. 4 Related work As mentioned above, a large number of layered agent architectures have been proposed for autonomous agent design and implementation. Chronologically, among the rst ones were e.g. NASREM [1] synchronous control of [3] SSS [5] TouringMachines [8], and ATLANTIS [12] Recently, a number of good overviews has appeared, e.g. chapter six of Arkin s textbook [2] pointing to the fact that the nature of the boundary between deliberation and reactive execution is not well understood at this time, leading to somewhat arbitrary architectural ....

[Article contains additional citation context not shown here]

Innes A. Ferguson. TouringMachines: An Architecture for Dynamic, Rational, Mobile Agents. PhD thesis, University of Cambridge, Computer Laboratory, Cambridge CB2 3QG, England, November 1992.

....in DIMA is provided with capabilities to reason on the agent state, on its acquaintances state and on the received performative and its various parameters. So, the communication module is cognitive. 3.2 The Agent Meta Behavior 3.2. 1 Issues Several hybrid architectures were proposed (see [13] and [22] to build agents out of two or more components which can be either reactive or cognitive. The reactive components are given some kind of precedence over the cognitive ones and the scheduling of these several components is static. Moreover, each component can be viewed as invoking a ....

....reactive and cognitive components. Decision Perception Figure 4: A static scheduler Let s consider a simple economic agent with two behaviors: a perception and a decision components. Figure 4 gives an example of scheduling that is implicitly used in most existing architectures (see for example [13]) The agent activity may be summarized within a perception and then decision loop. To underline the previous problem, we give the following scenario: At time t, Firm 1 has a model of its competitive Firm 2 which has an above price and a lower quality. Firm 1 activates its decision process. ....

[Article contains additional citation context not shown here]

I. A. Ferguson. TouringMachines: An Architecture for Dynamic, Rational, Mobile Agents. PhD thesis, University of Cambridge, Clare Hal, 1992.

....often own a deliberation behavior, and a communication behavior and or a perception behavior. Moreover, the use of a modular approach facilitates the integration of new modules such as a learning module. 3.2 The Agent Meta Behavior 3.2. 1 Issues Several hybrid architectures were proposed (see [4] and [9] to build agents out of two or more components which can be either reactive or cognitive. The reactive components are given some kind of precedence over the cognitive ones and the scheduling of these several components is static. Moreover, each component can be viewed as invoking a ....

....and cognitive components. Decision Perception Figure 5: A static scheduler Let s consider a simple economic agent with two behaviors: a perception and a decision components. Figure 5 gives an example of static scheduling that is implicitly used in most existing architectures (see for example [4]) The agent activity may be summarized within a perception and then decision loop. 8 To underline this problem of control, we give the following scenario: At time t, Firm 1 has a model of its competitive Firm 2 which has a greater price and a lower quality. Firm 1 activates its decision ....

[Article contains additional citation context not shown here]

I. A. Ferguson. TouringMachines: An Architecture for Dynamic, Rational, Mobile Agents. PhD thesis, University of Cambridge, Clare Hall, 1992.

.... fault diagnosis systems) Recent agent architectures that provide signi cantly support for mentalistic notions include JAM [ Huber, 1999 ] Agent0 [ Shoham, 1993 ] or PLACA [ Thomas, 1995 ] In addition to these monolithic architectures, a number of layered approaches such as TouringMachines [ Ferguson, 1992 ] or InteRRaP [ Mueller, 1996 ] exist, that provide different modules for reactive behavior, planning, or scheduling. Many of these systems are built on the foundations of BDI logics. Within the MAS community, the BDI model [ Rao and George , 1991 ] has come to be possibly the best known and ....

A. Ferguson. TouringMachines: An Architecture for Dynamic, Rational Mobile Agents. PhD Thesis, University of Cambridge, Cambridge, USA, 1992.

....execution has been found. From these, activity spreads back downward to the behavior based layer, which is the only layer with direct access to the actoric functions de ned in the agent s world interface. This is an important di erence to the other layered architectures, such as those by Ferguson [10] and Kaebling [27] where there is concurrent access both to sensory input and actoric output, and where con icts among the layers have to be solved by applying appropriate global ltering and suppression mechanisms to ensure that a speci c layer only sees those parts of the input that are relevant ....

I. A. Ferguson, \TouringMachines: an architecture for dynamic, rational, mobile agents," Ph.D. thesis, Computer Science Lab., University of Cambridge, UK, 1992.

....agents react rapidly to asynchronous events without using complex reasoning. Many researchers have suggested that neither a completely reactive nor a completely cognitive approach is suitable for building complete solutions for real life applications. Hybrid models, such as TouringMachines [Ferguson 1992], InteRRaP [Mller and Pischel 1994] and the model presented in [Bussmann and Demazeau 1994] have been proposed to combine the advantages of both reactive and cognitive models. In these models, agents are decomposed in a set of modules which can in turn be of a reactive or cognitive nature. ....

I. A. Ferguson. TouringMachines: An Architecture for Dynamic, Rational, Mobile Agents. PhD thesis, Clare Hall, University of Cambridge, UK.

....one, which is capable of reacting to events that occur in the environment without engaging in complex reasoning. Often, the reactive component is given some kind of precedence over the deliberative one, so that it can provide a rapid response to important environmental events. TOURINGMACHINES [Fer92] INTERRAP [MPT95] and PRS [GL87] are instances of hybrid agent architectures. The details of the example architectures that have been mentioned in this section are beyond the scope of this thesis. However, a brief description of the PRS (Procedural Reasoning System) is included in section 3.1 ....

I. A. Ferguson. TouringMachines: An Architecture for Dynamic, Rational, Mobile Agents. Technical Report TR273, University of Cambridge, 1992.

.... details, see (Brazier, Eck, and Treur, 1997) Touring Machines, INTERRAP, ZEUS, and ADEPT In the remainder of this section it is discussed how the generic agent model GAM can be refined to obtain a formally specified design model for four other existing agent architectures: Touring Machines (Ferguson, 1992), INTERRAP (Mller, Pischel, and Thiel, 1995; Mller, 1996) ZEUS (Nwana, Ndumu and Lee, 1998) and ADEPT (Jennings, Faratin, Norman, O Brien, Wiegand, Voudouris, Alty, Miah, and Mamdani, 1996) The Touring Machines architecture described in (Ferguson, 1992) distinguishes three layers: a reactive ....

.... agent architectures: Touring Machines (Ferguson, 1992) INTERRAP (Mller, Pischel, and Thiel, 1995; Mller, 1996) ZEUS (Nwana, Ndumu and Lee, 1998) and ADEPT (Jennings, Faratin, Norman, O Brien, Wiegand, Voudouris, Alty, Miah, and Mamdani, 1996) The Touring Machines architecture described in (Ferguson, 1992) distinguishes three layers: a reactive layer, a planning layer, and a modelling layer; all layers process concurrently. The reactive layer can be formally specified as an instantiation of the the components world interaction management and agent interaction management in the generic agent model ....

Ferguson, I.A. (1992). Touring Machines: An Architecture for Dynamic, Rational, Mobile Agents. Ph.D. Thesis.

....system [26] is used for activating and deactivating sets of skills to create networks that change the state of the world and accomplish specific tasks ( 9] p. 238) The deliberation layer consists of the Adversarial Planner (AP) 23] Other examples of hybrid architectures are TOURINGMACHINES [24, 25] and Cypress [66] which combines a planner with a modification of the Procedural Reasoning System (PRS) 31] 3.2 Artificial Life: Situated Emotions In a survey of earlier artificial intelligence research on models of emotion by Pfeifer [48, 49] the majority of efforts were found to be ....

....based on propositional knowledge and beliefs. The inferences involved in causal attribution will be modelled at this level as well as the evaluation of one s actions in relation to norms or one s self ideal. Therefore, models of self and others will be placed at the conceptual level. Ferguson s [24, 25] use of Belief Desire Intention (BDI) models at the Modelling Layer of his hybrid TOURINGMACHINES architecture as well as Elliott s Concerns of Other database (see section 3.3) are relevant for the design of the conceptual level. ffl Action We plan to design and implement this component by ....

Ferguson I.A.: TouringMachines: An architecture for dynamic, rational, mobile agents, University of Cambridge, UK, PhD Thesis, 1992.

....control and monitor an agent s own behaviors are reused to monitor, interpret, and predict its partner s behaviors. Ferguson s work on the TouringMachine architecture focused on agents in dynamic, multi agent environments also involves an explicit modeling layer for tracking other agents[9]. A second area of related work is research specifically focused on agent modeling and plan recognition. Section 1 has discussed some of this work. In addition, some formal approaches for agent modeling, and in particular for recursive agent modeling, are also being investigated[11] Vidal and ....

I. A. Ferguson. TouringMachines: An architecture for dynamic, rational, mobile agents. PhD thesis, University of Cambridge, 1992.

.... 1992; Novak and Sigmund, 1992) which simply uses a memory span of one action to determine its next reaction) and less deep than others (e.g. many of the strategies used in Axelrod (1984) also planning architectures, such as APE (Spector and Hendler, 1990) and GOFER (Le Pape, 1990) see Ferguson (1992) for a detailed discussion) In addition, it provides an interesting view of the problem which is not so complex as to put substantial overheads onto an implementation (see chapter 7) Also, the notion of trust presented here is not intended to be the sole decision strategy in use in an agent at ....

....as we proceed in the following examples. 5.2 The Furniture Removers 5.2. 1 Why Furniture Removal The domain of furniture removal is a constrained example of what could, in future, be expected from embodied agents the sort of work which is heavy, messy, and potentially physically risky (Ferguson, 1992). For more physical risk, there is always work in nuclear installations, or in space. As an aside, enlightening presentations of this concept, albeit in idealised forms, are to be found in Isaac Asimov s work involving the robots. See in particular I Robot ) In addition the domain provides an ....

Ferguson, Innes A. 1992. TouringMachines: An Architecture for Dynamic, Rational, Mobile Agents. Tech. rept. 273. University of Cambridge Computer Laboratory.

....example priority based situation checking) have to be defined within the individual functions. One way to overcome this problem is the use of layered agent architectures, that have become an important direction in intelligent agent design over the past few years (see e.g. Brooks 86, Kaelbling 90, Ferguson 92, Firby 92, Lyons Hendriks 92, Dabija 93, Steiner et al. 93, Muller Pischel 94a, Muller Pischel 94c] Layering is a powerful concept for the design of resource bounded agents. It combines a modular structure with a clear control methodology, and supports a natural modelling of different ....

I. A. Ferguson. TouringMachines: An Architecture for Dynamic, Rational, Mobile Agents. PhD thesis, Computer Laboratory, University of Cambridge, UK,, 1992.

....A sub task of this layer, the continuous look ahead (or limited prediction) mech anism has been distinguished in this layer due to the assumed implementation strategy. Literally dozens of layered architectures have been recently proposed for autonomous system implementation (see e.g. [1, 2, 4, 6, 8, 28] for just some of them) Although there are differences both in the way of assigning various tasks to different layers and in the way the overall control of the system is executed, the general conclusion is that such layering is beneficial, if not necessary, in designing autonomous intelligent ....

....specification of system s reactions, and the highest layer is expected to perform path planning. However, in contrast to ours, his approach does not offer any tools for, nor even possibility of, analyzing the real time performance of the implemented system. Ferguson s TouringMachine architecture [6] is the most elaborate and complete of the three. It addresses both the real time aspects and higher cognitive functions (including manipulation of system goals, intentions and beliefs about itself and other agents) The main difference is in the perceptory and effectory paths: TouringMachine ....

Innes A. Ferguson. TouringMachines: An Architecture for Dynamic, Rational, Mobile Agents. PhD thesis, University of Cambridge, Computer Laboratory, Cambridge CB2 3QG, England, November 1992.

....situationally determined ; ii) that the agent has no global task constraints : which need to be reasoned about at run time; and (iii) that the agent s goal or desire system is capable of being represented implicitly in the agent s structure according to a fixed, pre compiled ranking scheme. (Ferguson, 1992a, pp29 30) Hybrid architectures, such as the prs, TouringMachines, InteRRaP, and cosy, are currently a very active area of work, and arguably have some advantages over both purely deliberative and purely reactive architectures. One potential difficulty with such architectures, however, is that they tend to ....

Ferguson, I. A. (1992a). TouringMachines: An Architecture for Dynamic, Rational, Mobile Agents.

.... Hanks 90, Russell Zilberstein 91, Kushmerick et al. 93, Boddy Dean 94] Architectures allowing agents to react to dynamics in execution while at the same time trying to achieve their longer term goals have been provided by research in planning and DAI (e.g. Cohen et al. 89, Firby 92, Ferguson 92, Lyons Hendriks 92, Muller Pischel 94a] ffl The task of a centrally maintaining and processing the knowledge about the shipping companies, their vehicles, and behaviour is very complex. Moreover, knowledge is often not even centrally available (real life company are not willing to share ....

I. A. Ferguson. TouringMachines: An Architecture for Dynamic, Rational, Mobile Agents. PhD thesis, Computer Laboratory, University of Cambridge, UK,, 1992.

....introduces the different modes of interaction within this hierarchy. An application of the AHA architecture is discussed in section 4. Finally, section 5 gives conclusions and outlines future work. 2 Related Work Our work on hierarchical agents continues the tradition of layered computation [3] [6] [2] The combination of a layered approach and a belief, desire, and intention (BDI) architecture found in the InteRRaP system [7] 11] most closely resembles our approach. Like [7] we take a pragmatic approach to the problem of designing an agent and do not propose a new theory for knowledge ....

I. A. Ferguson. TouringMachines: An Architecture for Dynamic, Rational, Mobile Agents. PhD thesis, University of Cambridge, UK, 1992.

....the different modes of interaction within this hierarchy. Examples from animation of virtual worlds are given in section 4. Finally, section 5 gives conclusions and outlines future work. 2 Related Work Our work on hierarchical agents continues the tradition of layered computation [Brooks, 1986] [Ferguson, 1992] [ Bonasso et al. 1995 ] The combination of a layered approach and a belief, desire, and intention (BDI) architecture found in the InteRRaP system [Fischer et al. 1995] Muller et al. 1994 ] most closely resembles our approach. Like [ Fischer et al. 1995 ] we take a pragmatic approach to ....

I. A. Ferguson. TouringMachines: An Architecture for Dynamic, Rational, Mobile Agents. PhD thesis, University of Cambridge, UK, 1992.

....whereas layered architectures use different representations and algorithms implemented in separate layers to perform these functions. Most uniform architectures, for example PRS [10] do not make any specific commitments on how reaction and deliberation should be interleaved. In addition, Ferguson [5] argues that There are a number of other reasons for advocating a layered control approach, including increased behavioural robustness and operational concurrency, as well as improved program comprehensibility and system testability and analysability. p.48) However, as pointed out by Wooldridge ....

I.A. Ferguson. TouringMachines: An Architecture for Dynamic, Rational, Mobile Agents. PhD thesis, University of Cambridge, 1992.

.... Furthermore, the notion of emerging makes the resulting behaviour of an agent hard to understand, even harder to predict, and almost impossible to define and control [24] In the past few years, layered architecture has been presented for supporting the integration of deliberation and reaction AI [11, 12, 18, 24]. The main idea is to construct the functionalities of an agent into two or more hierarchically organized layers that interact with each other in order to achieve coherent behaviour of the agent as a whole. While lower reactive layer monitors the outside world, the higher layer can do some complex ....

I.A. Ferguson. Touring Machines: an Architecture for Dynamic, Rational, Mobile Agents. PhD Thesis, Computer Laboratory, University of Cambridge, UK, 1992.

.... RearAgentThreshold then change velocity(Observer, RearAvoidanceVelocity) where FrontalAgentThreshold, FrontalAvoidanceVelocity, RearAgentThreshold, and RearAvoidanceVelocity are parameters associated with 4 Due to space restrictions much detail will, in fact, be omitted and presented elsewhere [Fer92]. 8 Figure 3: Appropriate situation action rules can enable an agent to avoid obstacles and kerbs, prevent it from straying over lane markings, and, when no other events require attention, adjust the agent s orientation to one of four orthogonal directions (0 ffi , 90 ffi , 180 ffi , or ....

Innes A. Ferguson. TouringMachines: An Architecture for Dynamic, Rational, Mobile Agents. PhD thesis, Computer Laboratory, University of Cambridge, Cambridge, UK, 1992. Forthcoming.

....or office environment. The tool makes use of several modes of input and output, as well as employing a number of interface and communications media with which to present information and interconnect geographically distributed system users. The People Finder employs agent based design techniques [Fer92] as a way of integrating a variety of heterogeneous applications running on a number of different computing platforms and operating systems. Specifically, the PeopleFinder s functionality is carried out by a collection of coordinated software agents which can: i) assume user delegated tasks ....

.... Unix) and facilitates the interconnection of system users via telephone, email, and voice messaging (see Figure 3) The agents used in the PeopleFinder application are based on the CALVIN open agent framework (see Figure 4) which in turn is an extension of the TouringMachine architecture [Fer92]. This framework provides application developers with a powerful set of agent programming tools including libraries of intra and inter agent protocols (e.g. KQML 2 ) sensory and effectory apparatus, internal behavior APIs, persistent storage management, and (currently under consideration) ....

Innes A. Ferguson. TouringMachines: An architecture for dynamic, rational, mobile agents. Ph.D. thesis, Computer Laboratory, University of Cambridge, Cambridge UK, 1992.

.... which users are logged on to the various multi platform computer networks throughout the building) The agents used in the PeopleFinder application are based on the CALVIN open agent framework (Ferguson and Davlouros 1995) which in turn is an extension of the TouringMachine architecture (Ferguson 1992). 1 This framework provides application developers with a powerful set of agent programming tools 1 The Communicating Agents Living Vicariously In Networks (CALVIN) architecture is an agent frame including libraries of intra and inter agent protocols (e.g. KQML 2 ) sensory and effectory ....

Ferguson, Innes A. 1992. TouringMachines: An Architecture for Dynamic, Rational, Mobile Agents. Ph.D.

....or office environment. The tool makes use of several modes of input and output, as well as employing a number of interface and communications media with which to present information and interconnect geographically distributed system users. The People Finder employs agentbased design techniques [Fer92] as a way of integrating a variety of heterogeneous applications running on a number of different computing platforms and operating systems. Specifically, the PeopleFinder s functionality is carried out by a collection of coordinated software agents which can: i) assume user delegated tasks ....

.... Unix) and facilitates the interconnection of system users via telephone, email, and voice messaging (see Figure 3) The agents used in the PeopleFinder application are based on the CALVIN open agent framework (see Figure 4) which in turn is an extension of the TouringMachine architecture [Fer92]. This framework provides application developers with a powerful set of agent programming tools including libraries of intra and inter agent protocols (e.g. KQML 2 ) sensory and effectory apparatus, internal behavior APIs, persistent storage management, and (currently under consideration) ....

Ferguson, Innes A. TouringMachines: An Architecture for Dynamic, Rational, Mobile Agents. Ph.D. diss., Computer Laboratory, University of Cambridge, Cambridge UK, 1992.

....provide submit maintain PS profiles enhance the robustness of the system with respect to the uncertainties and dynamics of the environment. The SIGMA market model described above has been implemented as a collection of specialized CALVIN agents. An extension of the TouringMachine architecture (Ferguson 1992), the CALVIN agent framework is an open architecture for facilitating the development of highly concurrent, embedded agent oriented applications. As such, the framework provides application developers with a powerful set of agent programming tools including libraries of intra and inter agent ....

Ferguson, I.A. 1992. TouringMachines: An architecture for dynamic, rational, mobile agents. Ph.D. Thesis, Computer Laboratory, University of Cambridge, Cambridge, UK.

....model of practical rationality which is centered around its capability for reasoning abductively about other agents mental states. Introduction This paper is concerned with aspects of the design and implementation of an integrated agent control architecture, the TouringMachine architecture (Ferguson 1992; Ferguson 1995) suitable for controlling and coordinating the actions of an autonomous rational agent embedded in a partiallystructured, dynamic, multi agent world. Implemented as a number of concurrently operating, latency bounded, task achieving control layers, the TouringMachine architecture ....

....control framework s mediation enables each layer to examine data from other layers, inject new data into them, or even remove data from the layers. This has the effect of altering, when required, the normal flow of data in the affected layer(s) So, in the road driving domain of the TouringWorld (Ferguson 1992) for example, the reactive rule in layer R to prevent an agent from straying over lane markings can, with the appropriate control rule present, be overridden should the agent embark on a plan to overtake the agent in front of it. Inputs to and outputs from layers are generated in a synchronous ....

[Article contains additional citation context not shown here]

Ferguson, I.A. 1992. TouringMachines: An architecture for dynamic, rational, mobile agents. Ph.D. thesis, Computer Laboratory, University of Cambridge, Cambridge, UK.

....having to cope with multi agent interactions, unpredictability, uncertainty, resource constrained tasks, and real time environmental change. While it is true that a number of simplifications have been in order to analyze more easily the behavior of TouringMachines (these are detailed elsewhere (Ferguson, 1992, pages132 134) it is fair to say that the TouringWorld is a reasonably faithful approximation to certain classes of real world, multi agent domains; for example, automated factory office floors or road traffic environments. This article is concerned with the design and implementation of an ....

.... is, upon characterizing those aspects of the intended real time road navigation domain that would most significantly constrain the TouringMachine agent design and after due consideration of the requirements for producing autonomous, effective, robust, and flexible behaviors in such a domain (Ferguson, 1992, pages 26 32) the TouringMachine architecture has been designed through integrating a number of reactive and suitably designed deliberative control functions. Implemented as a number of concurrently operating, latency bounded, task achieving control layers, the resulting TouringMachine ....

[Article contains additional citation context not shown here]

Ferguson, I.A. 1992. TouringMachines: An Architecture for Dynamic, Rational, Mobile Agents. Ph.D. thesis, Computer Laboratory, University of Cambridge, Cambridge UK.

....incidents, and recover dynamically from poor decisions. All of this, of course, on top of accomplishing whatever tasks it was originally assigned to do. This paper is concerned with the design and implementation of a novel integrated agent control architecture, the TouringMachine architecture [5 7], suitable for controlling and coordinating the actions of autonomous rational agents embedded in a partially structured, dynamic, multi agent world. Upon carrying out an analysis of the intended TouringMachine task domain that is, upon characterizing those aspects of the intended real time ....

.... each conflict, rightsof way protocols in operation, as well as any environmental and physical situational constraints (for example, the presence of other entities) or motivational forces (for example, an agent s own internal goals) that may constrain the possible actions that the agent can take [7]. 4 Experimenting with TouringMachines The research presented here adopts a fairly pragmatic approach toward understanding how complex environments might constrain the design of agents, and, conversely, how different task constraints and functional capabilities within agents might combine to ....

[Article contains additional citation context not shown here]

Innes A. Ferguson. TouringMachines: An Architecture for Dynamic, Rational, Mobile Agents. Ph.D. thesis, Computer Laboratory, University of Cambridge, Cambridge UK, 1992.

....incidents, and recover dynamically from poor decisions. All of this, of course, on top of accomplishing whatever tasks it was originally assigned to do. This paper is concerned with the design and implementation of a novel integrated agent control architecture, the TouringMachine architecture [Fer91, Fer92a, Fer92b, Fer92c], suitable for controlling and coordinating the actions of autonomous rational agents embedded in a partially structured, dynamic, multi agent world. Upon carrying out an analysis of the intended TouringMachine task domain that is, upon characterizing those aspects of the intended real time ....

.... urgency of each conflict, rights of way protocols in operation, as well as any environmental and physical situational constraints (e.g. the presence of other entities) or motivational forces (e.g. an agent s own internal goals) that may constrain the possible actions that the agent can take [Fer92c]. 4 Experimenting with TouringMachines The research presented here adopts a fairly pragmatic approach toward understanding how complex environments might constrain the design of agents, and, conversely, how different task constraints and functional capabilities within agents might combine to ....

[Article contains additional citation context not shown here]

Innes A. Ferguson. Touring-Machines: An Architecture for Dynamic, Rational, Mobile Agents. Ph.D. thesis, Computer Laboratory, University of Cambridge, Cambridge UK, 1992.

No context found.

I. A. Ferguson. TouringMachines: An Architecture for Dynamic, Rational, Mobile Agents. Phd thesis, University of Cambridge, 1992.

No context found.

I.A. Ferguson. TouringMachines: An Architecture for Dynamic, Rational, Mobile Agents. PhD thesis, Clare Hall, University of Cambridge, 1992.

No context found.

Ferguson, I. A. (1992). TouringMachines: An Architecture for Dynamic, Rational, Mobile Agents. PhD thesis, University of Cambridge.

No context found.

I. A. Ferguson. TouringMachines: An Architecture for Dynamic, Rational, Mobile Agents. Clare Hall, University of Cambridge, Cambridge, UK, 1992.

No context found.

I. Ferguson, Touring Machines: An architecture for Dynamic, Rational Agents, Ph.D. Dissertation, University of Cambridge, UK, 1992.

No context found.

I. A. Ferguson. Touring Machines: An Architecture for Dynamic, Rational, Mobile Agents. Ph.D. Thesis, computer laboratory, Univ. of Cambridge, UK. 1992.

No context found.

Ferguson, I. A. 1992. Touring Machines: An Architecture for Dynamic, Rational, Mobile Agents. Ph.D. Dissertation, Computer Laboratory, University of Cambridge.

No context found.

Ferguson, I. A. 1992. Touring Machines: An Architecture for Dynamic, Rational, Mobile Agents. Ph.D. Dissertation, Computer Laboratory, University of Cambridge.

No context found.

Ferguson, I. A. (1992b), TouringMachines: An Architecture for Dynamic, Rational, Mobile Agents, PhD Thesis, Computer Laboratory, University of Cambridge, UK.

No context found.

Ferguson, I.A. (1992) Touring Machines: An architecture for dynamic, rational, mobile agents. PhD thesis, Technical Report No. 273, University of Cambridge, Computer Laboratory, 1992.

No context found.

Ferguson, I. A. 1992. Touring Machines: An Architecture for Dynamic, Rational, Mobile Agents. Ph.D. Dissertation, Computer Laboratory, University of Cambridge.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC