Chaining backward

In backward chaining, the conclusion is known in advance, but the reasoning that is needed to reach that conclusion is not. This sounds a little odd. How can we know what we are going to conclude and yet not know how to reach that conclusion In fact, it is common for us to know where we want to go, but not how to get there we have a hypothesis, which is to be proved or disproved, or a goal that is to be attained, but are unclear about what steps are needed to reach that goal. Just like forward chaining, backward chaining 

In a backward-chaining tree, routes from an oval node, working down from the top, are drawn as unconnected lines if either route can be taken this corresponds to an OR condition  

The two lines are joined by an arc if both steps must be taken. 

IF spend a little THEN plead poverty AND promise to do better next year  

Suppose that, as our wedding anniversary approaches, our goal is to avoid having to return home to live with mother. Assuming that we are spineless enough that we need to consult an expert system for advice on how this catastrophe can be avoided, the goal is defined as  

---

In backward chaining, the inference engine would search through the knowledge base until it found a rule, the conclusion of which was "Go to live with mother." It would find three such rules ... 

From such a woolly starting point, time may be wasted while the ES tries to find out what the user really wants. On the other hand, if there are many different rules that lead to the action , it is possible that an ES using backward chaining will pursue many fruitless chains before reaching a satisfactory conclusion. 

Whether it follows forward or backward chaining, the system will pose a series of questions that allow it to gather more information about the problem until it successfully answers the user s query or runs out of alternatives and admits defeat. Just like the way that the system chooses to respond to a query, the exact sequence of steps that it takes is not preprogrammed into the knowledge base, but is determined "on the fly" by the scheduler, which reassesses the options after each response that the user provides. 

Inference paradigms. The basic inference mechanisms of forward-chaining and backward-chaining must be integrated into a real-time execution environment. 

Specific rules are based on the attributes of congeners, including the physiochemical parameters used to determine the QSAR equation, the biological activity, and the molecular structure. Backward chaining, using these rules with specific instances of substituents, yields detailed shape and character for the receptor model. For instance, an abstracted specific rule may take the form ... 

The above examples used to illustrate the specific rules for backward chaining are similar to other attempts at receptor mapping. ) However, these other methods do not account for interactions that may be based on a combination of effects such as hydrophobicity and ligand potency. For instance, a rule that might apply to a compound with a substituted phenyl ring may take the form( ... 

The disconnection approach [5] is adopted in this work because it is amenable to backward chaining systems. The starting point is the target compound, which is, in this case, a Diels-Alder product. The target compound is broken or disconnected into two distinct parts called synthons. The synthons are the ideal representations of the actual reactants used to produce the target compound. Synthons embody the physical properties of the actual compounds they represent. 

A sample inference mechanism using these facts (given the natural backward chaining of Prolog) might be... 

SpinPro is a typical backward chaining, rule-based expert system. Rule-based systems are systems in which the expert s knowledge is encoded primarily in the form of if-then rules, i.e., if a set of conditions are found to be true then draw a conclusion or perform an action. "Backward chaining" refers to the procedure for finding a solution to a problem. In a backward chaining system, the inference engine works backwards from a hypothesized solution to find facts that support the hypothesis. Alternative hypotheses are tried until one is found that is supported by the facts. 

In backward chaining the rules are evaluated starting from the conclusion (i.e., server is connected) and finding out the necessary condition. In first iteration the access must be granted (rule 4) to connect to the server. To grant access, the server must be online — one of the given assertions — and the user must be an administrator (rule 2). Since the online criterion is one of the given assertions, rule 3 holds the condition to be fulfilled. 

---