Software Agents: An Overview
`
`Hyacinth S. Nwana
`
`Intelligent Systems Research
`
`Advanced Applications & Technology Department
`BT Laboratories, Martlesham Heath
`
`Ipswich, Suffolk, IP5 7RE, U.K.
`
`e-mail: hyacinth@info.bt.co.uk
`
`Tel: (+44 1 473) 605457
`
`fax: (+441473) 642459
`
`Knowledge Engineering Review, Vol. 11, No 3, pp. 205-244, October/November 1996.
`
`© Cambridge University Press, 1996
`
`Abstract
`
`Agent software is a rapidly developing area of research. However, the overuse of the word
`‘agent’ has tended to mask the fact that, in reality, there is a truly heterogeneous body of
`research being carried out under this banner. This overview paper presents a typology of
`agents. Next, it places agents in context, defines them and then goes on, inter alia, to
`overview critically the rationales, hypotheses, goals, challenges and state-of-the-art
`demonstrators of the various agent types in our typology. Hence, it attempts to make explicit
`much of what is usually implicit in the agents literature. It also proceeds to overview some
`other general issues which pertain to all the types of agents in the typology. This paper
`largely reviews software agents, and it also contains some strong opinions that are not
`necessarily widely accepted by the agent community.
`
`EMC 1120
`
`
`
`
`
`
`
`
`1 Introduction
`
`The main goal of this paper is to overview the rapidly evolving area of software agents. The
`overuse of the word ‘agent’ has tended to mask the fact that, in reality, there is a truly
`heterogeneous body of research being carried out under this banner. This paper places agents
`in context, defines them and then goes on, inter alia, to overview critically the rationales,
`hypotheses, goals, challenges and state-of-the-art demonstrators/prototypes of the various
`agent types currently under investigation. It also proceeds to overview some other general
`issues which pertain to all the classes of agents identified. Finally, it speculates as to the
`future of the agents research in the short, medium and long terms. This paper largely reviews
`software agents. Since we are overviewing a broad range of agent types in this paper, we do
`not provide a definition of agenthood at this juncture. We defer such issues until Section 4
`where we present our typology of agents.
`
`The breakdown of the paper is as follows. Section 2 notes the situation of smart agents
`research in the broad field of Distributed Artificial Intelligence (DAI) and provides a brief
`history. Section 3 identifies the scope of applicability of agents research and notes that there
`is a diverse range of interested parties. Before the core critical overview of the agent typology
`of Section 5, Section 4 provides our view of what smart agents are; it also identifies the
`different types of agents which fall under the ‘agents’ banner and warns that fruly smart or
`intelligent agents do not yet exist! They are still very much the aspiration of agent
`researchers. Section 6 overviews some more general issues on agents and and speculates
`briefly towards the future of agents in the short, medium and long terms. Section 7 concludes
`the paper.
`
`2 Software Agents: History and the Context of this Paper
`
`Software agents have evolved from multi-agent systems (MAS), which in turn form one of
`three broad areas which fall under DAI, the other two being Distributed Problem Solving
`(DPS) and Parallel Al (PAI). Hence, as with multi-agent systems, they inherit many of DAI’s
`motivations, goals and potential benefits. For example, thanks to distributed computing,
`software agents inherit DAI’s potential benefits including modularity, speed (due to
`parallelism) and reliability (due to redundancy). It also inherits those due to Al such as
`operation at the knowledge level, easier maintenance, reusability and platform independence
`(Huhns & Singh, 1994).
`
`The concept of an agent, in the context of this paper, can be traced back to the early days of
`research into DAI in the 1970s - indeed, to Carl Hewitt’s concurrent Actor model (Hewitt,
`1977). In this model, Hewitt proposed the concept of a self-contained, interactive and
`concurrently-executing object which he termed ‘actor’. This object had some encapsulated
`internal state and could respond to messages from other similar objects: an actor
`
`“is a computational agent which has a mail address and a behaviour. Actors communicate by
`message-passing and carry out their actions concurrently” (Hewitt, 1977, p. 131).
`
`Broadly, for the purposes of this paper, we split the research on agents into two main strands:
`the first spanning the period 1977 to the current day, and the second from 1990 to the current
`day too. Strand 1 work on smart agents, which begun in the late 1970s and all through the
`1980s to the current day, concentrated mainly on deliberative-type agents with symbolic
`internal models; later in this paper, we type these as collaborative agents. A deliberative agent
`is
`
`
`
`
`
`
`
`
`“one that possesses an explicitly represented, symbolic model of the world, and in which
`decisions (for example about what actions to perform) are made via symbolic reasoning”
`(Wooldridge, 1995, p. 42).
`
`Initially, strand 1 “work concentrated on macro issues such as the interaction and
`communication between agents, the decomposition and distribution of tasks, coordination and
`cooperation, conflict resolution via negotiation, etc. Their goal was to specify, analyse, design
`and integrate systems comprising of multiple collaborative agents. These resulted in classic
`systems and work such as the actor model (Hewitt, 1977), MACE (Gasser ez al., 1987),
`DVMT (Lesser & Corkill, 1981), MICE (Durfee & Montgomery, 1989), MCS (Doran ef al.,
`1990), the contract network coordination approach (Smith, 1980; Davis & Smith, 1983),
`MAS/DAI planning and game theories (Rosenschein, 1985; Zlotkin & Rosenschein, 1989;
`Rosenschein & Zlotkin, 1994). These ‘macro’ aspects of agents as Gasser (1991) terms them,
`emphasises the society of agents over individual agents, while micro issues relate specifically
`to the latter. In any case, such issues are well summarised in Chaib-draa ez al. (1992), Bond &
`Gasser (1988) and Gasser & Huhns (1989). More recent work under this strand include
`TAEMS (Decker & Lesser, 1993; Decker, 1995) DRESUN (Carver ef al., 1991; Carver &
`Lesser, 1995), VDT (Levitt et al., 1994), ARCHON (Wittig, 1992; Jennings ef al., 1995),
`Note that game theoretic work should arguably not be classed as a macro approach; it may,
`indeed, lie more towards the micro end of the spectrum.
`
`In addition to the macro issues, strand 1 work has also been characterised by research and
`development into theoretical, architectural and language issues. In fact, such works evolve
`naturally, though not exclusively, from the investigation of the macro issues. This is well
`summarised in Wooldridge & Jennings (1995a), and in the edited collections of papers:
`Wooldridge & Jennings (1995b) and Wooldridge ez al. (1996).
`
`However, since 1990 or thereabouts, there has evidently been another distinct strand to the
`research and development work on software agents - the range of agent types being
`investigated is now much broader. Thus, this paper is complementary to Wooldridge &
`Jennings’ (1995a) by placing emphasis on this strand although, naturally, there is some
`overlap, i.e. it overviews the broadening typology of agents being investigated by agent
`researchers. Some cynics may argue that this strand arises because everybody is now calling
`everything an agent, thereby resulting, inevitably, in such broadness. We sympathise with this
`viewpoint; indeed, it is a key motivation for this paper - to overview the extensive work that
`goes under the ‘agent’ banner. Essentially, our point is that in addition to investigating macro
`issues and others such as theories, architectures, languages, there has also been an
`unmistakable trend towards the investigation of a broader range of agent types or classes. The
`context of this paper is summarised in Table 1 below.
`
`Strand 1 Macro issues Bond & Gasser (1988)
`Gasser & Huhns (1989)
`Chaib-draa et al. (1992)
`Gasser et al. (1995)
`
`
`
`
`
`
`
`
`
`Theories, architectures & Wooldridge & Jennings
`languages (1995a, 1995b)
`
`Wooldridge et al. (1996)
`
`Strand 2 Diversification in the types of | This paper covers this!
`agents being investigated
`
`Table 1 - Brief History of Software Agents and the Context of this Paper
`
`3 Who are Investigating Software Agents for What and Why?
`
`We eschew answering this question in a futuristic sense in favour of providing a flavour of
`the scope of the research and development underway in universities and industrial
`organisations. The range of firms and universities actively pursuing agent technology is quite
`broad and the list is ever growing. It includes small non-household names (e.g. Icon, Edify
`and Verity), medium-size organisations (e.g. Carnegie Mellon University (CMU), General
`Magic, Massachusetts Institute of Technology (MIT), the University of London) and the real
`big multinationals (e.g. Alcatel, Apple, AT&T, BT, Daimler-Benz, DEC, HP, IBM, Lotus,
`Microsoft, Oracle, Sharp). Clearly, these companies are by no means completely
`homogeneous, particularly if others such as Reuters and Dow Jones are appended to this list.
`
`The scope of the applications being investigated and/or developed is arguably more
`impressive: it really does range from the mundane (strictly speaking, not agent applications)
`to the moderately ‘smart’. Lotus, for example, will be providing a scripting language in their
`forthcoming version of Notes which would allow users to write their own individual scripts in
`order to manage their e-mails, calendars, and set up meetings, etc. This is based on the view
`that most people do not really need ‘smart’ agents. Towards the smart end of the spectrum are
`the likes of Sycara’s (1995) visitor hosting system at CMU. In this system, “task-specific”
`and “information-specific” agents cooperate in order to create and manage a visitor’s
`schedule to CMU. To achieve this, first, the agents access other on-line information resources
`in order to determine the visitor’s areas of interest, name and organisation and resolve the
`inevitable inconsistencies and ambiguities. More information is later gamered including the
`visitor’s status in her organisation and projects she is working on. Second, using the
`information gathered on the visitor, they retrieve information (e.g. rank, telephone number
`and e-mail address) from personnel databases in order to determine appropriate attendees (i.e.
`faculty). Third, the visitor hosting agent selects an initial list of faculty to be contacted,
`composes messages which it dispatches to the calendar agents of these faculties, asking
`whether they are willing to meet this visitor and at what time. If the faculty does not have a
`calendar agent, an e-mail is composed and despatched. Fourth, the responses are collated.
`Fifth, the visitor hosting agent creates the schedule for the visitor which involves booking
`rooms for the various appointments with faculty members. Naturally, the system interacts
`with the human organiser and seeks her confirmation, refutations, suggestions and advice.
`
`Most would agree that this demonstrator is pretty smart, but its ‘smartness’ derives from the
`fact that the ‘value’ gained from individual stand-alone agents coordinating their actions by
`working in cooperation, is greater than that gained from any individual agent. This is where
`agents really come into their element.
`
`More examples of applications are described later but application domains in which agent
`solutions are being applied to or investigated include workflow management, network
`management, air-traffic control, business process re-engineering, data mining, information
`
`4
`
`
`
`
`
`
`
`
`retrieval/management, electronic commerce, education, personal digital assistants (PDAs), e-
`mail, digital libraries, command and control, smart databases, scheduling/diary management,
`etc. Indeed, as Guilfoyle (1995) notes
`
`“in 10 years time most new IT development will be affected, and many consumer products
`will contain embedded agent-based systems”.
`
`The potential of agent technology has been much hailed, e.g. a 1994 report of Ovum’s, a UK-
`based market research company, is titled “Intelligent agents: the new revolution in software”
`(Ovum, 1994). The same firm has apparently predicted that the market sector totals for agent
`software and products for USA and Europe will be worth at least $3.9 billion by the year
`2000 in contrast to an estimated 1995 figure of $476 million (computed from figures quoted
`in Guilfoyle, 1995). Such predictions are perhaps overly optimistic.
`
`Moreover, as King (1995) notes telecommunications companies like BT and AT&T are
`working towards incorporating smart agents into their vast networks; entertainment, e.g.
`television, and retail firms would like to exploit agents to capture our program viewing and
`buying patterns respectively; computer firms are building the software and hardware tools
`and interfaces which would harbour numerous agents; Reinhardt (1994) reports that IBM
`plans (or may have already done) to launch a system, the IBM Communications Systems
`(ICS), which would use agents to deliver messages to mobile users in the form they want it,
`be it fax, speech or text, depending on the equipment the user is carrying at the time, e.g. a
`PDA, a portable PC or a mobile phone. At BT Laboratories, we have also carried out some
`agent-related research on a similar idea where the message could be routed to the nearest
`local device, which may or may not belong to the intended recipient of the message. In this
`case, the recipient’s agent negotiates with other agents for permission to use their facilities,
`and takes into consideration issues such as costs and bandwidth in such negotiations (Titmuss
`et al., 1996). At MIT, Pattie Maes’ group is investigating agents that can match buyers to
`sellers or which can build coalitions of people with similar interests. They are also drawing
`from biological evolution theory to implement demonstrators in which some user only
`possesses the ‘fittest” agents: agents would ‘reproduce’ and only the fittest of them will
`survive to serve their masters; the weaker ones would be purged.
`
`It is important to note that most of these are still demonstrators only: converting them into
`real usable applications would provide even greater challenges, some of which have been
`anticipated but, currently, many are unforeseen. The essential message of this section is that
`agents are here to stay, not least because of their diversity, their wide range of applicability
`and the broad spectrum of companies investing in them. As we move further and further into
`the information age, any information-based organisation which does not invest in agent
`technology may be committing commercial hara-kiri.
`
`4 What is an agent?
`
`We have as much chance of agreeing on a consensus definition for the word ‘agent’ as Al
`researchers have of arriving at one for ‘artificial intelligence’ itself - nil! Recent postings to
`the software agents mailing list (agents@sunlabs.eng.Sun.COM) prove this. Indeed, in a
`couple of these postings, some propounded the introduction of a financial and/or legal aspect
`to the definition of agents, much to the derision of others. There are at least two reasons why
`it is so difficult to define precisely what agents are. Firstly, agent researchers do not ‘own’
`this term in the same way as fuzzy logicians/Al researchers, for example, own the term ‘fuzzy
`logic’ - it is one that is used widely in everyday parlance as in travel agents, estate agents, etc.
`
`
`
`
`
`
`
`
`Secondly, even within the software fraterity, the word ‘agent’ is really an umbrella term for
`a heterogeneous body of research and development. The response of some agent researchers
`to this lack of definition has been to invent yet some more synonyms, and it is arguable if
`these solve anything or just further add to the confusion. So we now have synonyms
`including knowbots (i.e. knowledge-based robots), softbots (software robot), taskbots (task-
`based robots), userbots, robots, personal agents, autonomous agents and personal assistants.
`To be fair, there are some good reasons for having such synonyms. Firstly, agents come in
`many physical guises: for example, those that inhabit the physical world, some factory say,
`are called robots; those that inhabit vast computer networks are sometimes referred to as
`softbots; those that perform specific tasks are sometimes called taskbots; and autonomous
`agents refer typically to mobile agents or robots which operate in dynamic and uncertain
`environments. Secondly, agents can play many roles, hence personal assistants or knowbots,
`which have expert knowledge in some specific domain. Furthermore, due to the multiplicity
`of roles that agents can play, there is now a plethora of adjectives which precede the word
`‘agent’, as in the following drawn only from King’s (1995) paper: search agents, report
`agents, presentation agents, navigation agents, role-playing agents, management agents,
`search and retrieval agents, domain-specific agents, development agents, analysis and design
`agents, testing agents, packaging agents and help agents. King’s paper is futuristic and
`provides a role-specific classification of agents, and so such rampant metaphorical use of the
`word is fine. But there is also another view that it gives currency to others to refer to just
`about anything as an agent. For example, he considers “print monitors for open printing, fax
`redial, and others” (p. 18) as agents, albeit simple ones. As Wayner & Joch (1995) write,
`somewhat facetiously,
`
`“the metaphor has become so pervasive that we’re waiting for some enterprising company to
`advertise its computer switches as empowerment agents™ (p. 95).
`
`We tend to use the word slightly more carefully and selectively as we explain later.
`
`When we really have to, we define an agent as referring to a component of software and/or
`hardware which is capable of acting exactingly in order to accomplish tasks on behalf of its
`user. Given a choice, we would rather say it is an umbrella term, meta-term or class, which
`covers a range of other more specific agent types, and then go on to list and define what these
`other agent types are. This way, we reduce the chances of getting into the usual prolonged
`philosophical and sterile arguments which usually proceed the former definition, when any
`old software is conceivably recastable as agent-based software.
`
`4.1 A Typology of Agents
`
`This section attempts to place existing agents into different agent classes, i.e. its goal is to
`investigate a typology of agents. A typology refers to the study of types of entities. There are
`several dimensions to classify existing software agents.
`
`Firstly, agents may be classified by their mobility, i.e. by their ability to move around some
`network. This yields the classes of static or mobile agents.
`
`Secondly, they may be classed as either deliberative or reactive. Deliberative agents derive
`from the deliberative thinking paradigm: the agents possess an internal symbolic, reasoning
`model and they engage in planning and negotiation in order to achieve coordination with
`other agents. Work on reactive agents originate from research carried out by Brooks (1986)
`and Agre & Chapman (1987). These agents on the contrary do not have any internal,
`symbolic models of their environment, and they act using a stimulus/response type of
`
`
`
`
`
`
`
`
`behaviour by responding to the present state of the environment in which they are embedded
`(Ferber, 1994). Indeed, Brooks has argued that intelligent behaviour can be realised without
`the sort of explicit, symbolic representations of traditional Al (Brooks, 1991b).
`
`Thirdly, agents may be classified along several ideal and primary attributes which agents
`should exhibit. At BT Labs, we have identified a minimal list of three: autonomy, learning
`and cooperation. We appreciate that any such list is contentious, but it is no more or no less
`so than any other proposal. Hence, we are not claiming that this is a necessary or sufficient
`set. Autonomy refers to the principle that agents can operate on their own without the need for
`human guidance, even though this would sometimes be invaluable. Hence agents have
`individual internal states and goals, and they act in such a manner as to meet its goals on
`behalf of its user. A key element of their autonomy is their proactiveness, i.e. their ability to
`‘take the initiative’ rather than acting simply in response to their environment (Wooldridge &
`Jennings, 1995a). Cooperation with other agents is paramount: it is the raison d’étre for
`having multiple agents in the first place in contrast to having just one. In order to cooperate,
`agents need to possess a social ability, i.e. the ability to interact with other agents and
`possibly humans via some communication language (Wooldridge & Jennings, 1995a).
`Having said this, it is possible for agents to coordinate their actions without cooperation
`(Nwana ef al., 1996). Lastly, for agent systems to be truly ‘smart’, they would have to learn
`as they react and/or interact with their external environment. In our view, agents are (or
`should be) disembodied bits of ‘intelligence’. Though, we will not attempt to define what
`intelligence is, we maintain that a key attribute of any intelligent being is its ability to learn.
`The learning may also take the form of increased performance over time. We use these three
`minimal characteristics in Figure 1 to derive four types of agents to include in our typology:
`collaborative agents, collaborative learning agents, interface agents and truly smart agents.
`
`Smart
`Agents
`
`Collaborative
`Learning Agents
`
`Co0
`
`oS
`
`Autonomous
`
`Collaborative
`
`Agents Interface
`
`Agents
`
`Figure 1 - A Part View of an Agent Typology
`
`We emphasise that these distinctions are not definitive. For example, with collaborative
`agents, there is more emphasis on cooperation and autonomy than on learning; hence, we do
`not imply that collaborative agents never learn. Likewise, for interface agents, there is more
`emphasis on autonomy and learning than on cooperation. We do not consider anything else
`which lie outside the ‘intersecting areas’ to be agents. For example, most expert systems are
`largely ‘autonomous’ but, typically, they do not cooperate or learn. Ideally, in our view,
`agents should do all three equally well, but this is the aspiration rather than the reality. Truly
`smart agents do not yet exist: indeed, as Maes (1995a) notes “current commercially available
`agents barely justify the name”, yet alone the adjective ‘intelligent’. Foner (1993) is even
`more incandescent; though he wrote this in 1993, it still applies today:
`
`
`
`
`
`
`
`
`“... 1 find little justification for most of the commercial offerings that call themselves agents.
`Most of them tend to excessively anthromomorphize the software, and then conclude that it
`must be an agent because of that very anthropomorphization, while simultaneously failing to
`provide any sort of discourse or “social contract” between the user and the agent. Most are
`barely autonomous, unless a regularly-scheduled batch job counts. Many do not degrade
`gracefully, and therefore do not inspire enough trust to justify more than trivial delegation
`and its concomitant risks” (Foner, 1993, 39/40).
`
`In effect, like Foner, we assert that the arguments for most commercial offerings being agents
`suffer from the logical fallacy of petitio principii - they assume what they are trying to prove
`- or they are circular arguments. Indeed, this applies to other ‘agents’ in the literature.
`
`In principle, by combining the two constructs so far (i.e. static/mobile and
`reactive/deliberative) in conjunction with the agent types identified (i.e. collaborative agents,
`interface agents, etc.), we could have static deliberative collaborative agents, mobile reactive
`collaborative agents, static deliberative interface agents, mobile reactive interface agents,
`etc. But these categories, though quite a mouthful, may also be necessary to further classify
`existing agents. For example, Lashkari er al. (1994) presented a paper at AAAI on
`‘Collaborative interface agents’ which, in our classification, translates to static collaborative
`interface agents.
`
`Fourthly, agents may sometimes be classified by their roles (preferably, if the roles are major
`ones), e.g. world wide web (WWW) information agents. This category of agents usually
`exploits internet search engines such as WebCrawlers, Lycos and Spiders. Essentially, they
`help manage the vast amount of information in wide area networks like the internet. We refer
`to these class of agents in this paper as information or internet agents. Again, information
`agents may be static, mobile or deliberative. Clearly, it is also pointless making classes of
`other minor roles as in report agents, presentation agents, analysis and design agents, testing
`agents, packaging agents and help agents - or else, the list of classes will be large.
`
`Fifthly, we have also included the category of hybrid agents which combine of two or more
`agent philosophies in a single agent.
`
`There are other attributes of agents which we consider secondary to those already mentioned.
`For example, is an agent versatile (i.e. does it have many goals or does it engage in a variety
`of tasks)? Is an agent benevolent or non-helpful, antagonistic or altruistic? Does an agent lie
`knowingly or is it always truthful (this attribute is termed veracity)? Can you trust the agent
`enough to (risk) delegate tasks to it? Is it temporally continuous? Does it degrade gracefully
`in contrast to failing drastically at the boundaries? Perhaps unbelievably, some researchers
`are also attributing emotional attitudes to agents - do they get ‘fed up’ being asked to do the
`same thing time and time again? What role does emotion have in constructing believable
`agents (Bates, 1994)? Some agents are also imbued with mentalistic attitudes or notions such
`as beliefs, desires and intentions - referred to typically as BDI agents (Rao & Georgeff,
`1995). Such attributes as these provide for a stronger definition of agenthood.
`
`In essence, agents exist in a truly multi-dimensional space, which is why we have not used a
`two or three-dimensional matrix to classify them - this would be incomplete and inaccurate.
`However, for the sake of clarity of understanding, we have ‘collapsed’ this multi-dimensional
`space into a single list. In order to carry out such an audacious move, we have made use of
`our knowledge of the agents we know are currently ‘out there’ and what we wish to aspire to.
`Therefore, the ensuing list is to some degree arbitrary, but we believe these types cover most
`of the agent types being investigated currently. We have left out collaborative learning
`agents, see Figure 1, on the grounds that we do not know of the existence ‘out there’ of any
`
`
`
`
`
`
`
`
`such agents which collaborate and learn, but are not autonomous. Hence, we identify seven
`types of agents:
`
`. Collaborative agents
`
`. Interface agents
`
`. Mobile agents
`
`. Information/Internet agents
`. Reactive agents
`
`. Hybrid agents
`
`. Smart Agents
`
`There are some applications which combine agents from two or more of these categories, and
`we refer to these as heferogeneous agent systems. Such applications already exist even
`though they are relatively few. However, we also overview briefly such systems in the next
`section.
`
`Another issue of note (for completeness sake) is that agents need not be benevolent to one
`another. It is quite possible that agents may be in competition with one another, or perhaps
`quite antagonistic towards each other. However, we view competitive agents as potential
`subclasses of all these types. That is, it is possible to have competitive collaborative-type
`agents, competitive interface agents, competitive information agents, etc.
`
`4.2 A Critique of Our Typology
`
`As with our definition of agenthood, our typology of agents is bound to be contentious. Two
`official reviewers of this paper all took issue with it, but their suggestions are, in our opinion,
`either more debatable or unclear. One reviewer, reviewer 1, claimed that we have confused
`agents that are defined by what they do (information agents, interface agents and
`collaborative agents), and other types for the sort of technology that underpins these agents
`(mobile agents, reactive agents, hybrid agents). Thus, he/she would have preferred a 2-
`dimensional classification. The second reviewer mentioned a similar point but alluded to a
`different classification. To a large degree, we disagree with this criticism, though not fully.
`We believe we had already attempted, perhaps unsuccessfully, to pre-empt this criticism.
`Firstly, we would not group information agents, interface agents and collaborative agents in
`one large group: in our view, as we explained earlier, collaborative agents and interface
`agents are defined by what they are, while information agents are defined by what they do.
`Secondly, we do not agree fully with the assertion that mobile agents, reactive agents and
`hybrid agents are all underlying technologies for implementing the former classes. To this
`reviewer, interface agents are collaborative agents implemented using reactive technology!
`We simply disagree with this viewpoint. As we explain later in the paper, reactive agents for
`example, have a distinct philosophy, hypothesis, etc. which make it stand out from the rest.
`We have surveyed the area of technologies for building software agent systems in another
`paper, Nwana & Wooldridge (1996). However, we agree with the general thrust of the
`argument to some degree; for example, we fully accept the reviewers’ viewpoint that mobility
`is not a necessary condition for agenthood - a point which is implicit in Section 4.1, and
`which we explain later. Thirdly, we address such issues when we discuss the individual types
`more fully in the rest of the paper. Fourthly, we point out, explicitly, in Section 4.1 that
`agents exist in a truly multi-dimensional space, and that for the sake of clarity of
`
`
`
`
`
`
`
`
`understanding, we have collapsed this multi-dimensional space into a single list. To produce
`this list, we used a set of criteria which included inate properties of agents which we would
`prefer to see (autonomy, cooperation, learning), other constructs (static/mobile,
`deliberative/reactive), major roles (as in information agents) and whether they are hybrid or
`heterogencous. In a previous version of this paper where we had a more hierarchical
`breakdown, it turned out to be less clear. Fifthly, other typologies in the literature are equally
`as contentious. For example, Wooldridge & Jennings (1995a) broadly classify agents into the
`following: gopher agents, service performing agents and proactive agents. We believe this is
`too general and simplistic a classification. It is for these reasons that we opted for such a ‘flat®
`breakdown. To be fair, apart from the typology, these two reviewers were very
`complementary about the paper.
`
`In conclusion, our typology is not without its critics (but so are all others), but as reviewer 1
`pointed out “while T agree that most agents in the literature can be categorised into these
`types, I think the types are themselves faulty”. In this paper, we have deliberately traded in
`accuracy for clarity. Our typology highlights the key contexts in which the word ‘agent’ is
`used in the software literature.
`
`4.3 What Agents are Not
`
`In general, we have already noted that a software component which does not fall in one of the
`intersecting areas of Figure 1 does not count as an agent. In any case, before the word ‘agent’
`came into vogue within the computing/Al fraternity, Minsky, in his Society of Mind (1985),
`had already used it to formulate his theory of human intelligence. However, Minsky used it to
`refer to much more basic entities:
`
`«. to explain the mind, we have to show how minds are built from mindless stuff, from parts
`that are much smaller and simpler than anything we’d consider smart... But what could those
`simpler particles be - the “agents” that compose our minds? This is the subject of our
`book...” (Minsky, 1985, 18).
`
`Clearly, Minsky’s use of the word ‘agent’ is quite distinct from its use in this paper.
`
`Furthermore, as noted earlier, expert systems do not meet the preconditions of agenthood, as
`do most knowledge-based system applications. Modules in distributed computing
`applications do not constitute agents either as Huhns & Singh (1994) explain. First, such
`modules are rarely ‘smart’, and hence much less robust than agents are (or should be); they
`also do not degrade gracefully. Second, in agent-based systems generally, the communication
`involves involves high-level mess