LP-guided Search for Constraint Satisfaction Problems
نویسنده
چکیده
Although addressing many of the same problems, the fields of Artificial Intelligence (AI) and Operations Research (OR) were seemingly developed independent of each other. Only recently has an effort been made to collaborate between the two in order to design smarter and faster algorithms. By merging the mathematical programming techniques of OR and the inference and/or randomized methods used in AI research, an entirely new (and more powerful) set of tools becomes available. Traditionally, however, it has been difficult to integrate these different techniques in a combinatorial setting, where variables take on purely discrete domains. This is largely due to the fact that Linear Programming was designed to optimize systems with linear constraints. Since there is a significant amount of overhead involved in formulating and solving an LP in the discrete setting, the challenge lies in identifying the point at which the information retrieved from an LP is more expensive than the traditional search procedure. In this paper, I develop a framework for using Linear Programming relaxations of constraint satisfaction problems (CSPs) to guide local, combinatorial search. In particular, I evaluate this technique on the Quasigroup Completion Problem (QCP), which is discussed in Section 3. I compare the performance of pure CSP techniques with a pure Linear Programming solution, and finally a hybrid of the two.
منابع مشابه
Partial Constraint Satisfaction Problems and Guided Local Search
A largely unexplored aspect of Constraint Satisfaction Problem (CSP) is that of overconstrained instances for which no solution exists that satisfies all the constraints. In these problems, mentioned in the literature as Partial Constraint Satisfaction Problems (PCSPs), we are often looking for solutions which violate the minimum number of constraints. In more realistic settings, constraints vi...
متن کاملApproximations and Randomization to Boost CSP Techniques
In recent years we have seen an increasing interest in combining constraint satisfaction problem (CSP) formulations and linear programming (LP) based techniques for solving hard computational problems. While considerable progress has been made in the integration of these techniques for solving problems that exhibit a mixture of linear and combinatorial constraints, it has been surprisingly diff...
متن کاملDistributed Guided Local Search for Solving Binary DisCSPs
We introduce the Distributed Guided Local Search (DistGLS) algorithm for solving Distributed Constraint Satisfaction Problems. Our algorithm is based on the centralised Guided Local Search algorithm, which is extended with additional heuristics in order to enhance its efficiency in distributed scenarios. We discuss the strategies we use for dealing with local optima in the search for solutions ...
متن کاملAn Empirical Study of Multi-Point Constructive Search for Constraint Satisfaction
Multi-Point Constructive Search (MPCS) is a constructive search technique which borrows the idea from local search of being guided by multiple viewpoints. MPCS consists of a series of resource-limited backtracking searches: each starting from an empty solution or guided by one of a set of high quality, “elite” solutions encountered earlier in the search. This paper focuses on MPCS as applied to...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 2005