Much ado has been made about the inclusion of “rules engines” within modern BPM suites, but all rules engines aren’t created equal. Some rules engines are similar in concept to what used to be commonly called “expert systems” and are as a result both more powerful and more complex than the rules engines that are simply made up of externalized functions or code. I’ll simplify the concepts and explain why you may want to take the type of rules engine a given BPM suite uses into account when making a purchasing decision.
A note on terminology usage: Within this article, I’ll use “expert systems” to describe the more powerful and more complex rules engines, whereas I’ll use “procedural rules” to describe the more lightweight, procedural rules engines.
The definition captured in the first line here (source: Wikipedia) is an excellent overview of what differentiates expert systems from procedural rules engines. Simply put, expert systems are designed to consider a set of information much like a human, selecting from a set of possible outcomes simultaneously, whereas the procedural rules engines are told to simply consider one possible outcome at a time. The benefit to the former is an ability to think in a less structured manner, thereby encouraging a larger universe of possible outcomes, whereas the benefit of the latter is simplicity.
BPM and Rules Engines: Real-world deployment
Here’s a real-world example… Say you are a modern quantitative hedge fund, and you want to consider a bevy of possible securities trading strategies given pricing and fundamental data for a given industry’s equities. Those possible strategies could number in the tens or hundreds (or more). If you are using a more powerful rules engine, you would “insert one or more facts” into the rules engine, and it would fire all of the rules that apply, yielding a set of results or decisions. The rules engine thinks more like a human and considers all of the possibilities that have been coded into it. If you’re using a less complex, procedural rules engine, you would have to code in the execution of each individual rule (usually structured using an IF… THEN – or similar – construct), possibly making for much bulkier code. However, on the flip side, your logic is simpler and easier to understand for those who aren’t used to working with expert systems.
Like with all other decisions regarding which system you should choose for your business, the decision comes down to your specific business needs. Are you a quantitative hedge fund like the theoretical one described in the previous paragraph, and do you have a large set of possible trading strategies to consider for each set of pricing data that you receive? If so, you likely want to heavily weigh whether a candidate BPMS has a full-featured, embedded expert system. If, on the other hand, you are an online retail company that has a limited number of possible discount and shipping strategies, the use of an expert system may be a negative due to the additional complexity and your staff’s learning curve associated with learning to think in a manner necessary to develop a set of rules for an expert system.
In summary, all “rules engines” are not created equal. Carefully consider the relative strengths and weaknesses of your candidate BPMS vendors’ “rules engines” – whether they are expert systems or procedural rules engines – and ensure that you weigh that information according to your specific business needs. No matter what your decision, Princeton Blue, Inc., can help you make the right decision and implement your BPM solution in the most cost-effective manner; please visit our website at www.princetonblue.com and request us to contact you or contact us at (908) 369-0961.