A. J. Davenport and E.P.K. Tsang. Solving constraint satisfaction sequencing problems by iterative repair. In Proceedings, 14th UK Planning and Scheduling Special Interest Group Workshop, Colchester, UK, November 1995.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....instances. However, on some other instances, it cannot reduce domains enough to make complete search tractable. Hence, di erent incomplete approaches have been proposed, that leave out exhaustivity, trying to quickly nd approximately optimal solutions in an opportunistic way, e.g. local search [1, 7, 8], genetic algorithms [14] or ant colony optimization (ACO) approaches [11] This paper describes and compares three incomplete approaches for the car sequencing problem. Section 2 introduces the car sequencing problem, for which several greedy heuristics are described and examined in section 3. ....

....will inevitably be exceeded; an utilization rate close to 0 indicates that the demand is very low with respect to the capacity of the station. 2. 3 Test Suites All considered instances are available in the benchmark library CSPLib [6] The rst test suite contains 70 problem instances, used in [1, 7, 11] and grouped into 7 sets of 10 instances per utilization rate 0.60, 0.65, 0.70, 0.75, 0.80, 0.85, and 0.90; e.g. the instances within the group 0.70 are named 70 01, 70 10, and we refer to the group as 70 . All these instances are feasible ones, and have 200 cars to sequence, 5 options, ....

[Article contains additional citation context not shown here]

A. Davenport and E. P. K. Tsang. Solving constraint satisfaction sequencing problems by iterative repair. In Proceedings of the First International Conference on the Practical Applications of Constraint Technologies and Logic Programming, 345 { 357, 1999

....and Olovsson. 1.3. 4 Scheduling and planning Scheduling and planning problems are especially suited to formulation in constraint satisfaction terms, see for example work by Kreuger et al. 29, 28] Fox [10] Sadeh et al. 45, 46] Le Pape [39] Smith et al. 55, 54, 56] Davenport and Tsang [6], El Sakkout and Wallace [47] and Kamarainen [27] Often, scheduling problems are formulated with specialized constraints that encompass the functionality of a large set of basic constraints as described in the articles by Beldiceanu [1] and R egin [41, 42] Here we present several problems ....

Davenport, A., and Tsang, E. Solving constraint satisfaction sequencing problems by iterative repair.

....different options. The utilization percentage of an option i corresponds to the ratio of the number m i of cars requiring option i with respect to the maximum number of cars in a sequence which could have option i while satisfying the capacity constraint on i (i.e. 100 m i q i = n p i ) [5]. A high utilization percentage indicates that the demand is very close to the capacity. Hence, 19] and [17] introduced value ordering heuristics: the idea is to assign first the cars requiring options with high utilization percentages, corresponding to critical options. Such value ordering ....

....actually optimal, even though Ant P solver cannot be used to prove this optimality. 4 Related works Global vs local constraints: Generally speaking, to solve a CSP that contains permutation constraints, one had better to use a global constraint which usually handles it more efficiently. Indeed, [5] shows that using such global constraints always reduces the search space, even though it can be slower to actually solve the problem (this case usually happens on easy instances) The advantage of using global constraints is well illustrated by the all interval series problem. Indeed, 14] ....

[Article contains additional citation context not shown here]

A.J. Davenport and E.P.K. Tsang, `Solving constraint satisfaction sequencing problems by iterative repair', in Proceedings of the first international conference on the practical applications of constraint technologies and logic programming (PACLP), pp. 345--357, (1999).

....a specialized arcconsistency propagator. The constraints in the problem are all converted to global cardinality constraints and the propagator described in [16] is applied to these constraints. Several local search techniques have been applied to the Car Sequencing problem. Davenport and Tsang [8] defined a new class of problems called the Constraint Satisfaction Sequencing Problem, which is essentially a CSP where the variables all have the same domain values and an all different constraint is defined over the variables. They solved the problem using hill climbing with a variation of the ....

A. Davenport and E. Tsang. Solving constraint satisfaction sequencing problems by iterative repair. In Proceedings of the First International Conference on the Practical Application of Constraint Technologies and Logic Programming, pages 345--357, London, 1999.

....et al. 10] and R egin and Puget [7] solve this version of the problem using backtracking search with specialized propagators to maintain arc consistency during the search. Local search techniques have also been developed for this version of the problem including a hill climbing approach [3] and a simulated annealing approach [9] However, while this specification has demand and capacity constraints, it omits time window, change over, and balancing constraints important in our version of the problem. More directly related is the work done by ILOG on the vehicle sequencing problem ....

A. Davenport and E. Tsang. Solving constraint satisfaction sequencing problems by iterative repair. In Proc. of PACLP-99, pages 345--357, 1999.

....et al. 10] and R egin and Puget [7] solve this version of the problem using backtracking search with specialized propagators to maintain arc consistency during the search. Local search techniques have also been developed for this version of the problem including a hill climbing approach [3] and a simulated annealing approach [9] However, while this speci cation has demand and capacity constraints, it omits time window, change over, and balancing constraints important in our version of the problem. More directly related is the work done by ILOG on the vehicle sequencing problem for ....

A. Davenport and E. Tsang. Solving constraint satisfaction sequencing problems by iterative repair. In Proc. of PACLP-99, pages 345-357, 1999.

....for general non binary constraints, but the system developed by Davenport and Tsang is the version used in this paper. There has also been work on adapting it to solve optimisation problems [1, 17] GENET has also tackled the real world industrial problem of car sequencing on production lines [4]. The network is set up to represent a CSP by having each label (variable value pairing) denoted by a node. The label nodes corresponding to a particular variable form a cluster. Each label node can be in an on or o state, but only one node in each cluster may be on. This on state for a label ....

A. Davenport and E. P. K. Tsang. Solving constraint satisfaction sequencing problems by iterative repair: an application to car sequencing. In The Practical Application of Constraint Technologies and Logic Programming Conference - PACLP99, pages 345-358. The Practical Application Company Ltd, 1999.

No context found.

A. J. Davenport and E.P.K. Tsang. Solving constraint satisfaction sequencing problems by iterative repair. In Proceedings, 14th UK Planning and Scheduling Special Interest Group Workshop, Colchester, UK, November 1995.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC