Exploiting Interchangeabilities for Case Adaptation

1999

Case adaptation is a complex problem for which no general method has been found. We consider the restricted domain of problems which can be formulated as constraint satisfaction, and propose a general method using dimensionality reduction based on constraint solving and interchangeability.

2003

In [1] we propose interchangeability based algorithms as methods for solving the case adaptation for the domain of problems which can be expressed as Constraint Satisfaction Problems. In this paper we extend the domain to Soft Constraint Satisfaction Problems and give generic adaptation methods based on soft interchangeability concepts. Many real-life problems require the use of preferences. This need motivates for the use of soft constrains which allows the use of preferences. We have defined interchangeability for soft CSPs in [2] by introducing two notions: (δ/α)substitutability/interchangeability and their algorithms. This paper presents howto build generic adaptation methods based on soft interchangeability. It gives an example of an application of a sales manager system for a car configuration domain and reports test results regarding number of (δ/α)interchangeability in random generated problems, thus number of adaptation alternatives.

1998

We address in this paper the adaptation of a case when a complete constraint model of the underlying problem is given. The idea is to apply methods from constraint based reasoning that allows the detection of "similar" solutions, which can be used to adapt a selected case to a new situation. We consider applications like configuration where a complete constraint model is available.

1995

Case adaptation, a central component of case-based reasoning, is often considered to be the most difficult part of a casebased reasoning system. The difficulties arise from the fact that adaptation often does not converge, especially if it is not done in a systematic way. This problem, sometimes termed the assimilation problem, is especially pronounced in the case-based design problem solving domain where a large set of constraints and features are processed. Furthermore, in the design domain, multiple cases must be considered in conjunction in order to solve the new problem, resulting in the difficulty of how to efficiently combine the cases into a global solution for the new problem.

The adaptation process is an important and complex step of case-based reasoning (CBR) and is most of the time designed for a specific application. This article presents a domain-independent algorithm for adaptation in CBR. Cases are mapped to a set of numerical descriptors filled with values and local constraint intervals. The algorithm computes every target solution descriptor by combining a source solution, a matching expressed as intervals of variations and dependencies between the source problem and its solution. It determines for every target solution descriptor an interval of the admissible values. In this interval, actual values satisfying global constraints can be chosen. This generic approach to adaptation is operational and introduces general and domain-independent adaptation operators. Therefore, this study is a contribution to the design of a general algorithm for adaptation in CBR. CBR uses past solved cases, called source cases, stored in a case base, in order to solve a new problem, called the target problem and denoted by ¦ ¨ § ©¦ .

2004

Combinatorial optimization is a powerful paradigm for representing complex problems. It has a wide range of applications such as planning, scheduling, resource sharing, in many domains such as transportation, production, mass marketing, network management, human resources management. Constraint satisfaction techniques provide efficient algorithms to prune search spaces.

Proc. of Third Workshop on Learning …, 2009

Constraint-based problems are hard combinatorial problems and are usually solved by heuristic search methods. In this paper, we consider applying a machine learning approach to improve the performance of these search-based solvers. We apply reinforcement learning in the context of Constraint Satisfaction Problems (CSP) to learn a value function, which results in a novel solving strategy. The motivation underlying this approach is to solve previously unsolvable instances.

1994

The retrieval of a suitable case is of crucial importance to the success of case-based reasoning. A good criterion for judging “case suitability” is how complex a case will be to adapt. However, it has proven difficult to directly calculate this measure of case “adaptability” without incurring the full cost of adaptation. This has led most researchers to exploit semantic similarity as a more tractable (albeit less accurate) answer to the question of case suitability.

Lecture Notes in Computer Science, 2002

Combinatorial optimization is a powerful paradigm for representing complex problems. It has a wide range of applications such as planning, scheduling, resource sharing, in many domains such as transportation, production, mass marketing, network management, human resources management. Constraint satisfaction techniques provide efficient algorithms to prune search spaces.

2001