问题
I am looking to integrate with a 'Generic Rules Engine' based on the request of a customer.
I think the objective is to allow business stakeholders to add 'Rules', and have those be incorporated into an overall metric calculated on a dataset. So far, the Rules i have heard seem like straightforward snippets of logic in the code. I suppose the drawback is that even though simple, this would still need to be coded... (as opposed to some kind of runtime or data driven rule specification automatically used in the analysis.)
hopefully not too vague - but anyone have any success with such a thing? which open source projects have the most promise?
thanks
回答1:
I have played around with DROOLS, a rule engine from JBOSS. I have seen it use in large scale production systems. It offers representation of rules in various different formats such as -- flat rule file written in JAVA or MVEL; using DROOLS rule flow, and decision tables composed in EXCEL.
The execution of rules are using RETE algorithm, which is supposedly faster due to rule memorization and variable sharing. As pointed out by Doug, there are a lot of information on Wikipedia
回答2:
Expert Systems were the AI rage in the 80s.
There's lots of info on Wikipedia on the Rete Algorithm
See also Inference Engine
One well regarded toolkit is CLIPS
来源：https://stackoverflow.com/questions/4118528/anyone-with-specific-pointers-or-experience-with-a-generic-rules-engine