Acerca de

Based on the article "Is it an Agent, or just a Program?: A Taxonomy for Autonomous Agents", written by Stan Franklin and Art Graesser. The meaning of agents in general are software program that respond to states and events in their environment without direct command from the agent's user or owner, but working on behalf of and in the interest of the owner. From the article, they already divided into several example of agents such as The MuBot Agent which carry meaning used to refer to two distinct ideas. The first is the agent's capacity to execute tasks independently. The agent's capacity to do domain-oriented reasoning is the second factor to consider, The AIMA Agent means an agent is defined as "anything that perceives its surroundings through sensors and acts on it through effectors" , The Maes Agent also means "Independent specialists are computational frameworks that in propensity a few complex energetic environment, sense and act independently in this environment, and by doing so realize a set of objectives or assignments for which they are outlined" and The SodaBot Agent means "Computer program specialists are programs that lock in in dialogs and arrange and facilitate exchange of data." But, the article already formalize them into one definition "an autonomous agents is a system that is positioned inside and a part of an environment that senses that environment and acts on it throughout time in pursuit of its own agenda and to affect what it detects in the future".

 

Furthermore, the article said that how are they different from other computer programs because back to the general definition an agents is an autonomous substance that exists in an environment and acts in a levelheaded way. Meanwhile a program on the other hand isn't as independent and cannot act normally, the program depends more exceedingly on the client of it. Based on the article said that "an autonomous agents has succeeded in recognizing between agents and programs". An agents require not be a program at all. It may be a robot or a school instructor. In the mean time computer program operators are by definition programs, but a program must degree up to a few marks to be an operator. Besides,  definition of an autonomous agents also moreover abdicate a lesson of operators so huge as not to guarantee extraordinary utility. 

​

Moreover, the article said that, the agents characteristics that differentiate from computer programs is agents can be effectively categorised based on which of these traits they possess. According to the writer definition, every agent meets the first four qualities. Other qualities are added to create potentially useful kinds of agents, such as mobile and learning agents. As a result, a hierarchical categorization based on set inclusion emerges. Mobile agents with learning capabilities are a subtype of mobile agents. Meanwhile, Writers divide software agents into categories based on the jobs they do, such as data gathering agents or email filtering agents. Alternatively, the writer might categorised them based on their control architecture. Sumpy would be a planning agent, whereas Etzioni and Weld's Softbot would be a fuzzy subsumption agent. Agents can also be classed based on the range and sensitivity of their senses, the scope and efficacy of their acts, or their internal condition. Furthermore, software agents are classified in three ways in Brustoloni's taxonomy such as regulation agents, planning agents, and adaptive agents. A regulatory agent, which is most often associated with temperature control via a thermostat or other forms of biological homeostasis, reacts to each sensory input as it arrives and always knows what to do. It doesn't plan or learn. Planning agents plan in the traditional AI sense (problem solving agents), the case-based paradigm (case-based agents), operations research-based methodologies (OR agents), or a variety of randomising algorithms (randomizing agent). The adaptive agents created by Brustoloni not only plan, but also learn.

​

​