Georg Ringwelski

Boosting distributed constraint satisfaction

Competition and cooperation can boost the performance of search. Both can be implemented with a portfolio of algorithms which run in parallel, give hints to each other and compete for being the first to finish and deliver the solution. In this paper we present a new generic framework for the application of algorithms for distributed constraint satisfaction which makes use of both cooperation and competition. This framework improves the performance of two different standard algorithms by one order of magnitude and can reduce the risk of poor performance by up to three orders of magnitude. Moreover it greatly reduces the classical idleness flaw usually observed in distributed hierarchy-based searches. We expect our new methods to be similarly beneficial for any tree-based distributed search and describe ways on how to incorporate them.

POOC: a platform for object-oriented constraint programming

In this paper, we describe an implementation-independent object-oriented interface for commercial and academic Constraint Solvers. This serves as a basis for evaluating different Constraint Solvers and for developing solver-independent applications. We show, how applications can use the interface, which solvers are already integrated into the framework and how additional solvers can be added. Furtermore, we provide to the community the described system as real Java packages via Internet, that even includes a basic but powerful Constraint Solver.

A SPACE-EFFICIENT BACKTRACK-FREE REPRESENTATION FOR CONSTRAINT SATISFACTION PROBLEMS

In this paper we present a radical approach to obtaining a backtrack-free representation for a constraint satisfaction problem: remove values that lead to dead-ends. This technique does not require additional space but has the drawback of removing solutions. We investigate a number of variations on the basic algorithm including the use of seed solutions, consistency techniques, and a variety of pruning heuristics. Our experimental results indicate that a significant proportion of the solutions to the original problem can be retained especially when an optimization algorithm that specifically searches for such “good” backtrack-free representations is employed. Further extensions increase solution retention by searching for high-coverage backtrack-free representations, by removing tuples rather than values, and by combining multiple backtrack-free representations. Our approach elucidates, for the first time, a three-way trade-off between space complexity, potential backtracks, and solution loss and enables algorithms that can actively reason about the trade-off between space, backtracks, and solution loss.

Distributed Constraint Satisfaction with Cooperating Asynchronous Solvers

A Constraint Satisfaction Problem (CSP) is to find an assignment to a set of variables that is consistent wrt. a set of constraints over these variables. CSPs frequentlyarise in applications of distributed artificial intelligence [3] and may often not be solved bya centralized constraint solver for privacyor security reasons. In this distributed case (DCSP) constraints and variables are distributed among multiple automated agents.

An Arc-Consistency Algorithm for Dynamic and Distributed Constraint Satisfaction Problems

This paper presents the new DDAC4 algorithm for dynamic arc consistency enforcement in distributed constraint satisfaction problems (CSP). The algorithm is an adaptation of the well-known AC-4 algorithm to system settings where constraints can be added and deleted in concurrent processes. It is the first algorithm for arc-consistency enforcement in this system setting. Arc-consistency is achieved whenever the overall system reaches quiescence after a constraint is added or deleted.

Object-Oriented Constraint Programming with J.CP

We give a short introduction to Asynchronous Constraint Solving (ACS), which is a new execution model to solve constraint satisfaction problems through constraint propagation and search. Propagation is based on the theory of chaotic iteration and search is implemented with explicit constraint retraction. ACS is designed for the object-oriented development of distributed or strongly interactive applications. Our implementation J.CP makes use of this new execution model. J.CP is a Java package, which combines the declarativity of constraint programming with the features of the object-oriented programming language Java. Constraints are autonomous objects that can be posted and retracted in concurrently running solvers that communicate via common constraint variables.

Models for Human Computer Interaction in Scheduling Applications

There are many algorithms to solve scheduling problems, but in practice the knowledge of human experts almost always needs to be involved to get satisfiable solutions. However, human computer interaction in scheduling applications is often designed in a way, that does not leave much room for own decisions to the user. In this paper, we describe a set of decision support features that can be used to improve the human-computer-interaction in scheduling applications. Based on a study with 35 test subjects and overall 105 h of usability testing we verify that the use of the features improves both quality and practicability of the produced schedules.

Integrating Search Objects in Asynchronous Constraint Solving

Asynchronous Constraint Solving (ACS) integrates dynamic constraint processing into concurrent Object Oriented Programming. Cooperating constraint solvers run in parallel to the application program and infer actual variable domains incrementally from constraints that are added or retracted in the application thread. Constraint addition starts a chaotic iteration on the variable domains leading to a fixed point where no more domain reductions can be deduced from the constraint implementations. Constraint retraction removes all consequences of a constraint from the knowledge represented in the variables and can thus be considered the inverse operation to constraint addition.