Browse Prior Art Database

Rule based equation generation

IP.com Disclosure Number: IPCOM000208020D
Publication Date: 2011-Jun-20
Document File: 4 page(s) / 146K

Publishing Venue

The IP.com Prior Art Database

Abstract

This publication proposes a simple and generic extension of rule based languages in order to be able to express linear (integer or not) equations. This feature is useful to offer linear programming (LP) and mixed integer programming (MIP) modeling capabilities to users who are acquainted with rule based languages but not with regular mathematical programming modeling languages. It consists in adding the syntactic means of expression constraints, right hand sides and objectives in the right part of any rule.A portion of this article contains material which is subject to copyright protection. The copyright owner reserves all copyright rights.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 53% of the total text.

Page 01 of 4

Rule based equation generation

I. Technical Problem:

A. Context:

Writing linear programming (LP) or mixed integer programming (MIP) include 4 phases:

1- read input data,

2- generate a matrix for the LP/MIP engine

3- Solve de L/MIP problem (done by the optimization engine)

4- read engine solution and map it back to data, and output solution.

Present publication addresses phase 2 (generation phase), and creates the handles to address phase 4 too.

B. Problem to be solved:

The objective of this publication is to offer a user-friendly modeling language/environment, as well as a way to bypass re-design/re-code/re-deploy phases when changes need to be made in the mathematical model of a optimization based application.

II. Prior Art:

·MIP/LP modeling languages (http://en.

language#Algebraic ) address

matrix generation phase by allowing a mathematical-like syntax.
·Business Rules Management Systems (BRMS) allow user friendly natural-like languages, and bypass re-deployment phase of software life cycle

However, MIP/LP modeling languages need software engineers to
· learn a specific language (AMPL, GAMS, OPL etc..)
· Include initial modeling and modeling changes in the classical life cycle pattern

Further, BRMS does not address mathematical modeling.

III. Proposed solution: Description

The proposed solution consists in:

A. Use of BRMS technology to manage:

 Data transformation, from a any source (DBMS, POJO, WSDL etc..)
 Pseudo natural language to express rules, making it accessible to non-programmers  On the fly deployment of new constraints, or constraint modifications
 Replace iterative building of modeling expressions with intuitive rules

B. Add a set of constraint creation actions to be used as right part (consequence part) of business rules namely

1.Coefficient expression:

wikipedia

/Modeling

_

.org/

wiki

1


Page 02 of 4

Constraint(name,index1,..,indexn) += coefficient *

DecisionVariable(name, type (int/float), index1,…,indexn)

2.RHS (Right Hand Side) expression:

Constraint(name,index1,..,indexn)

value

Comparator in {

=,==}

3.Objective expression:

Objective(name,type(min/max)) += coefficient * DecisionVariable(name,

type (int/float), index1,…,indexn)

     Figures 1 to 3 below illustrate the step of the proposed solution. Figure 1 shows a flowchart for interpreting the constraint related action part of rules :

(This page contains 00 pictures or other non-t...