[rules-dev] Rules Complexity

Our clients needs an AML (Anti Money Lanudering) Monitoring system to be implemented. This system is suppose to monitor the customer, transaction and historical data and generate alerts based on the rules set by the clients. As we are in design and thought process of it, we thought of Drools for this. Our doubts are 1) Are we going in the right direction by using Drools? 2) Can we compare the current data and historical data and do a link analysis using Drools? 3) Can we represent sequential patterns of rules in Drools? 4) DRL or XML format of rules, which is the best method to represent rules? I tried writing complex rules using Drools, but not getting the flexibility. For eg: we need a rule as following: 1. Check the transactions of a non customer on the same day or short period (configurable) 2. Check if all the transactions' originator and beneficiary information are same or not. 3. If transaction amounts are of very small, calculate aggregate and find the sum crosses the threshold limit or not. If yes, raise the alert. Data Set is the list of non customers. From this list, while running the rule, how will I write the rule to find the aggregate for each of the customer and check the threshold? Warm Regards, Rajitha