On Combining Regression Analysis and Constraint Programming

Gervet, Carmen; Galichet, Sylvie

doi:10.1007/978-3-319-08855-6_29

Cited by 2 publications

(2 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…After enlarging the interval bounds of the input data we were able to find a solution with a 50 % split of traffic, but none with 40 − 60 or other combinations. This experimental study showed the strong impact of taking into account dependency constraints with simulations [13].…”

Section: Traffic Conservation Constraintsmentioning

confidence: 83%

El MUNDO: Embedding Measurement Uncertainty in Decision Making and Optimization

Gervet

Galichet

2015

European Project Space on Intelligent Systems, Pattern Recognition and Biomedical Systems

Self Cite

View full text Add to dashboard Cite

In this project we address the problem of modelling and solving constraint based problems permeated with data uncertainty, due to imprecise measurements or incomplete knowledge. It is commonly specified as bounded interval parameters in a constraint problem. For tractability reasons, existing approaches assume independence of the data, also called parameters. This assumption is safe as no solutions are lost, but can lead to large solution spaces, and a loss of the problem structure. In this paper we present two approaches we have investigated in the El MUNDO project, to handle data parameter dependencies effectively. The first one is generic whereas the second one focuses on a specific problem structure. The first approach combines two complementary paradigms, namely constraint programming and regression analysis, and identifies the relationships between potential solutions and parameter variations. The second approach identifies the context of matrix models and shows how dependency constraints over the data columns of such matrices can be modeled and handled very efficiently. Illustrations of both approaches and their benefits are shown.

show abstract

Section: Traffic Conservation Constraintsmentioning

confidence: 83%

El MUNDO: Embedding Measurement Uncertainty in Decision Making and Optimization

Gervet

Galichet

2015

European Project Space on Intelligent Systems, Pattern Recognition and Biomedical Systems

Self Cite

View full text Add to dashboard Cite

show abstract

“…The solution sets produced can be very large. This led to some research to extract the relationship between uncertain data that satisfy dependency constraints and possible solutions by applying regression analysis techniques [11]. The fuzzy and mixed CSP [9] coined the concept of uncontrollable variables, that can take a set of values but their domain is not meant to be pruned during problem solving (unlike decision variables).…”

Section: Related Workmentioning

confidence: 99%

Uncertain Data Dependency Constraints in Matrix Models

Gervet¹,

Galichet²

2015

Integration of AI and OR Techniques in Constraint Programming

View full text Add to dashboard Cite

Abstract. Uncertain data due to imprecise measurements is commonly specified as bounded intervals in a constraint decision or optimization problem. Dependencies do exist among such data, e.g. upper bound on the sum of uncertain production rates per machine, sum of traffic distribution ratios from a router over several links. For tractability reasons existing approaches in constraint programming or robust optimization frameworks assume independence of the data. This assumption is safe, but can lead to large solution spaces, and a loss of problem structure. Thus it cannot be overlooked. In this paper we identify the context of matrix models and show how data dependency constraints over the columns of such matrices can be modeled and handled efficiently in relationship with the decision variables. Matrix models are linear models whereby the matrix cells specify for instance, the duration of production per item, the production rates, or the wage costs, in applications such as production planning, economics, inventory management. Data imprecision applies to the cells of the matrix and the output vector. Our approach contributes the following results: 1) the identification of the context of matrix models with data constraints, 2) an efficient modeling approach of such constraints that suits solvers from multiple paradigms. An illustration of the approach and its benefits are shown on a production planning problem.

show abstract

On Combining Regression Analysis and Constraint Programming

Cited by 2 publications

References 8 publications

El MUNDO: Embedding Measurement Uncertainty in Decision Making and Optimization

El MUNDO: Embedding Measurement Uncertainty in Decision Making and Optimization

Uncertain Data Dependency Constraints in Matrix Models

Contact Info

Product

Resources

About