SOFTWARE AGENTS

BY

LEKE, OLUWADARA DEBORAH

2019/8155

SUBMITTED TO

DR TEMITOPE OGUNBIYI

NOVEMBER, 2023
TABLE OF CONTENTS

Introduction....................................................................................................................................1
Definition of Software Agents.......................................................................................................1
Difference between Software Agents and Expert Systems........................................................1
Types of Software Agents..............................................................................................................2
1 Collaborative Agents..........................................................................................................2
2 Interface Agents..................................................................................................................2
3 Mobile Agents......................................................................................................................3
4 Information/Internet Agents..............................................................................................3
5 Reactive Agents...................................................................................................................3
6 Hybrid Agents.....................................................................................................................4
Applications of Software Agents..................................................................................................4
1 Personal Information Management..................................................................................4
2 Electronic Commerce.........................................................................................................5
3 Business Process Management..........................................................................................7
Benefits of Software Agents..........................................................................................................7
Disadvantages of Software Agents...............................................................................................8
Introduction

Expert systems (ES) are one of the prominent research domains of AI but it’s development

techniques have several shortcomings. These ES shortcomings are exactly what Software agents

was developed to address. This is why it is important we talk about software agents when

looking at Expert Systems. (Guimaraes, 2006)

Definition of Software Agents

Sofware Agents are driven by goals, much as people do. They make autonomous judgments and

learn from experience, serving as a bridge between individuals and information. It draws from

limited rationality, AI, behavioral economics, control theory, and game theory. (Burr, Cristianini,

& Ladyman, 2018)

A software agent is a software object that operates autonomously on behalf of others,

demonstrating proactivity, responsiveness, learning, collaboration, and mobility, while the

programmer specifies their capabilities and required information. (Hurst, Nangle, Cunningham,

Somers, & Evans, 1997)

Difference between Software Agents and Expert Systems

According to Guimaraes (2006), These are the differences between Expert Systems and Software

Agents.

Expert System Software Agents

Provides tools to manage knowledge within Shares information, knowledge and tasks
specific domains among themselves and cooperate with each
other to achieve common goals.
They are not designed for reactive, proactive They have a flexible behavior, reactive,
behavior. proactive and social.
They are not coupled to their environments They are coupled to their environments
Types of Software Agents

According to Nwana (1996), There are seven types of agents namely

1 Collaborative Agents
Collaborative agents are static, large coarse-grained agents that emphasize autonomy and

cooperation with other agents to perform tasks for their owners. They may learn but it is

not a major focus. To have a coordinated set up, they may need to negotiate agreements.

The key general characteristics of these agents include autonomy, social ability,

responsiveness and proactiveness. The goal of collaborative agent systems is to create a

system that interconnects separately developed colaborative agents, enabling the

ensemble to function beyond their capabilities.

2 Interface Agents
Interface agents are personal assistants who work with users to complete tasks. They

support and aid consumers, frequently teaching them how to utilize certain programs. The

user's agent watches and monitors activities, learns shortcuts, and proposes more efficient

methods to do tasks. Interface agents learn to help users by observing, providing

feedback, giving specific instructions, and seeking peer guidance. They enhance their

performance by learning from the user, receiving feedback, and interacting with others.

Interface agents vary from collaborative agents in that they help users learn to use certain

apps (such as spreadsheets or operating systems) rather than needing explicit language.

Interface agents often seek advice from others rather than engage in bargaining, as

collaborative agents do. Learning modes are rote or parametric, with techniques like

evolutionary learning also being introduced.

3 Mobile Agents
 Mobile agents are software processes that roam wide area networks (WANs) like the

WWW, communicate with foreign hosts, gather information for their owner, then return

home after completing their assigned tasks. Duties may include booking flights and

monitoring telecommunications networks. However, movement is not essential nor

sufficient for agenthood. Mobile agents are independent and interact in a unique way,

unlike collaborative agents. Agents can collaborate and communicate by sharing internal

object and method locations with other agents. This allows agents to share data without

disclosing all of their information.

4 Information/Internet Agents
Information agents have come about because of the sheer demand for tools to help us

manage the explosive growth of information we are experiencing currently, and which

we will continue to experience henceforth. Information agents perform the role of

managing, manipulating or collating information from many distributed sources.

5 Reactive Agents
Reactive agents are simple, stimulus-response agents that do not have internal symbolic

models of their environment; instead they act/respond in a stimulus-response manner to

the present state of the environment in which they are embedded. They interact with other

agents in basic ways and perform task decomposition i.e. it is a collection of modules

which operate autonomously and are responsible for specific tasks (e.g. sensing, motor

control, computations, etc.). Communication between modules is minimal, and no global

model exists within any agent. Reactive agents operate on representations close to raw

sensor data.

Benefits of Reactive agents include:

 i. It is more robust and fault tolerant than other agent-based systems

ii. It is more adaptable and flexible than other software agents

iii. It doesn’t have a slow response time and brittleness like other AI systems

6 Hybrid Agents
The discussion over the best agent among the five categories of agents mentioned above

(collaborative, interface, mobile, internet, and reactive agents) is theoretical. Each kind

has its own set of strengths and limitations, thus enhancing strengths while eliminating

inadequacies is more important and is frequently done through a hybrid strategy.

Some applications combine agents from two or more categories and can be referred to as

Heterogeneous Agents

Applications of Software Agents

There are three promising areas using service-performing and predictive software agents which

include: Personal Information Management, Electronic commerce and Business Process

Management. Other areas not discussed include: messaging software, development tools,

information management, user interface software, process control, workflow management, and

network management. (Jennings & Wooldridge, 1996)

1 Personal Information Management

The rapid growth of network technology, particularly the widespread adoption of the Internet

by commercial and academic organizations, has exceeded the expectations of even computer

network professionals. By 1995, a significant portion of Internet traffic came from the World

Wide Web (WWW), a network application that did not exist in 1991. This surge in interest

has revolutionized academic research, allowing new findings to be disseminated almost

instantaneously via electronic means like the WWW, contrasting with the lengthy
dissemination times of the past. However, the abundance of information has introduced

challenges, such as difficulty in finding relevant data and managing the vast amounts of

information. An ideal solution would be a trained assistant to search, sift, and manage

information, a role that is impractical for humans due to cost. Instead, software agents are

being developed to fulfill this role. Pioneering work by Pattie Maes’ group at MIT has

produced prototypes like NEWT, an electronic news scanner that learns to identify articles of

interest to the user, acting as an extension of the user’s preferences and intentions.

2 Electronic Commerce

The Internet is increasingly used for buying and selling goods and services, beyond just

information repositories. Currently, consumers can order items like pizzas and CDs online,

and more expensive goods and services are expected to follow. In such environments,

software agents play a crucial role, with user agents representing consumers and business
agents representing suppliers. Users are also able to delegate tasks like buying cinema tickets

and booking restaurant reservations to their software agents. These agents will operate

autonomously, making decisions based on the user's preferences and interacting with other

service-providing agents in the electronic marketplace. For example, an agent might query

cinema agents for show times, then check restaurant agents for dining options that match the

user's preferences. If needed, it will reassess options to find the best match. This process

saves users time and effort by handling mundane tasks. Service-performing agents

demonstrate key attributes such as autonomy, social ability, responsiveness, and

proactiveness. They may also be mobile, moving between machines to reduce

communication costs, although this requires enhanced security measures to prevent risks

similar to computer viruses.

 3 Business Process Management
Software agents not only empower individuals but also enhance organizational efficiency by

managing complex business processes, such as aircraft design, shopping center construction,

or large contract tendering. These processes require coordination among various departments

like legal, technical, and marketing. Traditionally managed manually, these tasks often

resulted in inefficiencies and high costs. Limited computer support tools, like workflow

systems, have been introduced, but they are typically inflexible. Projects like ADEPT have

adopted a negotiating agents approach, where each department is represented by an

autonomous software agent capable of performing specific services. When a high-level task,

such as producing a tender, is delegated to the agent system, agents negotiate among

themselves to schedule and execute necessary services, forming binding contracts to ensure

timely completion. This approach preserves departmental autonomy and effectively

represents the distribution of data, control, and resources, providing a flexible and responsive

solution to managing business functions and optimizing resource use.

Benefits of Software Agents

Oniovosa, Valeen, & Abidemi (2016) states the following benefits of Software Agents.
 Agents make less work for the end user and application developer.

 The agent can adapt to its user preferences and habit over a course of time.

 It will intelligently get shared among the community.

 Mobile agents manage the users’ E-mail, fax, phone and pager as well as linking the user

to Telescript-enabled messaging and communication services such as America Online

and AT&T Personal Link Services.

  The most favorite area with respect to reactive agents is games and entertainment

industry.

 Shopping agents are ideal applications of Agent Builder agents. These agents can be used

to locate merchandise, compare prices, place orders, etc.

Disadvantages of Software Agents

The following according to Jennings and Wooldridge (1996) are the disadvantages of Software

Agents:

 Overhyped Expectations: The capabilities of software agents are often exaggerated,

leading to unrealistic expectations.

 Limited Problem-Solving: They cannot solve all complex problems in advanced

information processing systems, such as planning in uncertain environments, timely

responses to environmental changes, and accurately inferring user preferences.

 New Problems: The autonomy of software agents introduces new issues that need

addressing.

 Trust and Reliability: Users may be reluctant to trust autonomous agents to act on their

behalf.

 Potential for Errors: Agents, like email filters, may mistakenly delete important

messages, causing significant loss.

 Financial Risks: Agents making financial commitments on behalf of users can lead to

serious consequences if errors occur.

 REFERENCES
Burr, C., Cristianini, N., & Ladyman, J. (2018). An Analysis of the Interaction Between
Intelligent Software Agents and Human Users. Minds and Machines, 28(4), 735–774. DOI:
10.1007/s11023-018-9479-0
Guimaraes, T. (2006). Comparing Expert Systems and Agent Technology for KM. Emerging
Information Resources Management and Technologies. DOI: 10.4018/9781599042862.ch014
Hurst, L., Nangle, B., Cunningham, P., Somers, F., & Evans, R. (1997). Software Agents: A
review.
Jennings, N., & Wooldridge, M. (1996). Software agents. IEE Review, 42(1), 17–20.
DOI:10.1049/ir:19960101
Nwana, H. S. (1996). Software agents: An overview. The Knowledge Engineering Review, 11(3),
205–244. DOI: 10.1017/s026988890000789x
Oniovosa, O.-E., Valeen, O., & Abidemi, S. (2016). USE OF SOFTWARE AGENTS IN E-
COMMERCE: BENEFITS AND APPLICATIONS. International Journal of Advanced
Engineering and Management Research, 2(2456-3676).