Classification of Software Agents

Software agents can be classified on the basis of three orthogonal dimensions; these dimensions are as follows: 

1) Mobility: It refers to the degree to which an agent can move through networks. In some cases, the agent program executes only on the host system and does not move at all, it interacts with other systems only through the communication mechanism. In other cases, the agent programme can be moved to another computer and executed there.

 

In still others, agent programmes are capable of suspending their own execution on the current system, move their programmes and execution state to another computer, and 1esume execution there on the new system: 

i) Static Agents: These agents execute on systems in which they begin execution, ånd interact with other systems using communication mechanisms. They use embedded knowledge to assist in filtering and processing volumes of information.

 

ii) Mobile Agents: These agents are not bound to the system in which they start execution, and can, therefore, travel among other hosts in a network. They carry-out transactions without continuous, instructions from the user. Mobility is achieved by transportation of state and code to the new execution environment at the destination. Mobile agents are well-suited for e-commerce, as commercial transactions may require real time access to remote resources such as stock quotes and agent-agent negotiations.

 

2) Intelligence: It refers to the degree to which an agent can identify and perform tasks, in order to meet the objectives specified by users. Intelligence is a unique human trait. 

Human intelligence has several levels – the highest probably being creativity. But, on the other hand, the mind has learned to process constant sensory input signals at a trivial level so that the signals do not overwhelm us. These signals are classified into manageable sets of information by identifying the patterns, similarities, and differences. The sensory input is put in an appropriate compartment based on a set of learned rules. Each compartment has rules, for example, the human mind knows that if it has wings and feathers, it is a bird, if car races towards you, move to the side, if you want to the past tense of a verb not ending with "e" add "ed", otherwise add "d". get Now, all one needs to note is similarity, differences, and changes.

 

In case of exceptions, add the exceptions as well to the set rules.

3) Autonomy: This refers to the degree to which an agent can exercise control over its own actions and state In other words, autonomy refers to the agents' ability to act without supervision. In a typical user and computer program interaction paradigms, programs act only when users initiate them to do so. In an alternative to this paradigm, both users and computers can initiate and monitor events to meet a set objective.

 

Role of Software Agents for B2B EC 

1) The use of intelligent softwáre agents that automate the coordination activities among business partners is not only an advantage, but a necessity in the connected and networked world that companies operate in. 

2) Software (Intelligent) agents can be responsible for collecting and interpreting information of merchants and products, making decisions, even presenting payment information and settling transactions, thus reducing time and cost.

3) They will help buyers and suppliers achieve greater efficiencies in all of the important aspects of B2B, such as content, community and transaction. Additionally, agent based e-commerce happens on the Internet and does not require the high setup costs associated with EDI. 

4) Agents can help in automating all the repetitive activities in the following business aspects: 

i) Internal functions and activities, such as accounting, order fulfillment and personnel management;  

ii) External activities, such as supply chain management and e- procurement, which involve interaction with trading partners, B2B exchanges; and External customer centric activities such as eCRM, sales and services.