Modelling and Solving Fleet Assignment in a. Flexible Environment. Hani El Sakkout. IC-Parc, Imperial College, London SW7 2AZ.

Size: px
Start display at page:

Download "Modelling and Solving Fleet Assignment in a. Flexible Environment. Hani El Sakkout. IC-Parc, Imperial College, London SW7 2AZ."

Transcription

1 Modelling and Solving Fleet Assignment in a Flexible Environment Hani El Sakkout IC-Parc, Imperial College, London SW7 2AZ hhe@doc.ic.ac.uk Keywords Combinatorial Optimization, Constraint Satisfaction, Finite Domain Propagation, Constraint Repair, Fleet Assignment Abstract This paper reports on the modelling and solution of a large industrial combinatorial optimization problem using the ECL i PS e Constraint Logic Programming (CLP) platform. The problem is eet assignment: the assignment of a set of aircraft of dierent types to a predened shedule of ights. A single underlying model was developed and solved in ECL i PS e. Three optimization methods were applied in three stages of model development and solution. The three methods were used to meet dierent objectives. The rst was depth rst search branch and bound optimization in conjunction with Finite Domain propagation. This straightforward approach was used to develop and validate the model in the ECL i PS e platform. The second was mixed integer programming. The CLP model and solutions were exported to a mathematical solver and used to produce a globally optimal solution. Knowing the optimal solution is desirable, however obtaining good solutions close to the user's actual assignment is more useful in practice. The third method used a repair-based extension of the rst approach, and obtained high quality solutions close to the user's assignment. This multistage devolopment process resulted in the end user commissioning the development of an ECL i PS e -based application combining repair with Finite Domain propagation, with the purpose of improving their day to day eet assignment process. The exercise clearly reveals the exibility of the ECL i PS e platform in mapping a single model to dierent problem solving approaches, and illustrates how this exibility has important practical benets. 1

2 1 Introduction 1.1 Paper Outline This paper describes our experience in solving a large industrial eet assignment problem using the ECL i PS e Constraint Logic Programming platform. Three alternative solution strategies were used at dierent stages of the project to achieve dierent objectives. The British Airways (BA) shorthaul eet assignment requires the feasible allocation of aircraft of various types to a xed set of trips. Solutions to this problem must also satisfy a set of maintenance constraints and minimise a cost function. This large constraint satisfaction problem involves over 10,000 constraints and 8000 variables. The rst solution method involved depth rst search search in conjunction with a Finite Domain (FD) solver (Section 2). This approach was used to develop, verify and enhance a model for the problem. This model is formalised in Sections 1.3 and 1.4. In the second stage of the project, the model and solutions obtained in the rst stage were exported to a mixed integer solver. The globally optimal solution was obtained; Section 3 briey discusses this method. The global optimal is useful since it provides a measure against which other solutions are tested, but it is more useful in practice to have good quality solutions that minimise disruption to the user's current eet assignment. The third method used a repair hybrid based on the rst approach (Section 4). Knowledge of the problem structure obtained in the rst stage of the project was found to be invaluable. This method produced good quality solutions that were close to the user's current eet assignment. This multistage development process convinced the user to commission the development of a practical ECL i PS e -based repair hybrid based on the third method. The exercise demonstrates the exibility and practicality of the ECL i PS e platform in the modelling and solution of a large scale industrial problem. The implications of sharing the same model among a pure FD propagation implementation, mixed integer programming and a repairbased hybrid are discussed, and some conclusions are drawn in Section Background References The concept of arc consistency was introduced in [Mac77]; its combination with backtrack search was described in [HE80]; the notion of value propagation is due to [SS80]; the application of constraint methods to real arithmetic was surveyed in [Dav87]; and [Van89] describes in detail the integration of Finite Domain (FD) propagation methods into Logic Programming to produce FD Constraint Logic Programming (CLP) [Kow79, JL87]. ECL i PS e is a exible platform based originally on FD CLP and extended to facilitate incorporation of a number of other constraint solvers [ECR95].

3 1.3 The Fleet Assignment Problem Problem Description Fleet assignment involves the assignment of aircraft types to a xed set of ights. Each ight departs from a specied origin station at a specied time. The destination station is also xed, but the time of arrival is dependent on the aircraft type chosen. Naturally, the choice of aircraft type for a particular ight is limited to those available at the departure station at the time of departure. Requirements dictate that a certain percentage of each aircraft type must start and end the day at an airport capable of carrying out maintenance. The cost of each ight also depends on the type of aircraft chosen. The global cost function is dened as the sum of the ight costs. As explained below, detail of the user's original problem that may handled by a preprocessing stage has not been included in the model since it is not relevant to the search and optimization process The Problem Components A set of stations s 2 S The total number of aircraft overnighting at each station on(s) 2 N is given as part of the problem specication. The relation maint(s) holds i s is a maintenance station. A set of plane types p 2 P Each plane type is associated with a eet fs(p), and a required nightly maintenance percentage mpc(p). A set of ight legs l 2 L Each leg has an origin o(l) and a destination d(l), as well as a xed local departure time t dep (l) 2 T. A choice of assignments relation C : (L P ) This relation denes the space of possible assignment pairs (l 2 L; p 2 P ) of plane types to legs. Two functions are dened for each pair (l; p) 2 C : a cost c : LP! N, and the local time t ready : L P! T after which an aircraft of type p is ready for a subsequent departure from d(l) Feasibility and Optimality The task of solving the eet assignment problem is essentially to assign a particular plane type to each leg. Let each leg l 2 L be associated with a plane type variable pt l whose domain is dened as follows: dom(pt l ) = fp : (l; p) 2 Cg. Let each pair (s; p) be associated with an overnight variable on sp signifying the number of aircraft of plane type p overnighting at s, and the set of all such variables be ON.

4 A feasible solution is an assignment to all the plane type variables fpt l : l 2 Lg such that 9ON 8>< >: where delta(s; p; t) = (i) 8s : p2p on sp = on(s) (ii) P 8p : s2s on sp = fs(p) (iii) 8p : s2s^maint(s) on sp mpc(p) fs(p) (iv) 8s; p : jfl : o(l) = s ^ pt l = pgj = jfl : d(l) = s ^ pt l = pgj (v) 8s; p; t : on sp delta(s; p; t) l : pt l = p o(l) = s t dep (l) t? l : pt l = p d(l) = s t ready (l) t. An optimal solution is a feasible solution minimising the global cost function P l c(l; pt l ). 1.4 A Practical Model The problem formulation given above is not well suited to direct encoding of the constraints. Specically, constraints (iv) and (v) cannot be used for ecient constraint handling, as explained below. Constraint (iv) requires, for each station, the daily quantity of outgoing aircraft of each plane type to be equal to the daily quantity of incoming aircraft of the same type. Constraint (v) requires these quantities to be indexed to particular time points. To address this problem two sets of peripheral variables are dened: Boolean plane type variables, and Delta variables. This section introduces the more practical representation of the constraints that is common to all three methods presented here Boolean Plane Type Variables Since both constraints need a notion of plane type quantity, which is dependent on the assignment of plane type variables, it is necessary to introduce Boolean variables (in Binary representation). These are dened as follows: 8l; p : b lp = 1, pt l = p Constraints on quantities of aircraft of particular plane types are then dened in terms of these booleans directly Delta Variables In the more abstract specication above, constraint (v) as stated implies the computation of the dierence between departures and arrivals up to every time point in the day for each station and plane type pair (this value is designated delta(s; p; t) in the problem formulation), in order that the maximum may be identied. 1 An enhanced representation is more 1 It would be more precise to speak about ready times instead of arrival times. An aircraft must remain out of action for a minimum time in between an arrival and a subsequent departure.

5 practical and ecient; changes only occur to the delta(s; p; t) values after departure and arrival time-points from s. In addition, since the maximum is being computed, we need only observe the value of delta(s; p; t) at departure time-points ftj9l : t = t dep (l)^o(l) = sg. The value of delta(s; p; t) may also be constructed incrementally left-to-right on the time line, if the line is divided into adjacent intervals separated by ight departures. In other words the set of relevant time-points may be reduced to: MT (s) = ft start g [ ftj9l : t = t dep (l) ^ o(l) = sg [ ft end g where t start is an initial time point before which no arrivals or departures take place at s, and t end is a nal time point after which no arrivals or departures take place at s. Let I(s) be the set of time line intervals (t; t 0 ] such that t immediately precedes t 0 in an ordering of MT (s). We introduce delta variables of the form delta spt corresponding to the values for delta(s; p; t) for all t 2 MT (s)? ft start g. Constraints on these variables are dened for each station s, plane type p, and pair of adjacent intervals (t; t 0 ] and (t 0 ; t 00 ] in I(s): Constraint (v): delta spt 00 = P o(l)=s^t dep (l)=t 00 b lp? P d(l)=s^t 0 <t ready (l)t 00 b lp + delta spt 0 Constraint (iv): delta sptend = 0 2 The Pure FD Propagation Method 2.1 Motivation The constraint model given in Sections 1.3 and 1.4 is the product of an iterative process. Improvements to the user's original specication and data were brought about through the gradual removal of pre-processable computation, the correction of minor specication and data inconsistencies, and the subsequent incorporation of other reductions and enhancements as our understanding of the problem evolved. This model development and testing stage of the project made extensive use of the ECL i PS e \pure" Finite Domain propagation optimizer. The optimizer coupled depth rst branch and bound optimization with Finite Domain propagation. As the evolving model was tested on larger subsets of the data, the pure FD optimization process was made more ecient, primarily through the inclusion of redundant constraints, and the application of various search ordering strategies, as briey described below. 2.2 Enhancing Performance Redundant Constraints If chosen well, redundant constraints may produce signicant improvements in the behaviour of the Finite Domain solver. The earlier pruning they typically give rise to will reduce

6 the number of nodes searched as well as the number of backtracks. General Delta Constraints These redundant constraints are produced by performing total ow reasoning on all incoming and outgoing aircraft at a station. This is independent of plane type, unlike the ow reasoning of constraint (v). This allows us to relate the expected total number of aircraft on the ground at departure time-points with the total numbers of each plane type. By linking the new delta(s; t) variables with the existing delta(s; p; t) variables the two forms of reasoning interact. The General Delta Constraints are redundant in that, if ow logic is consistent for each plane type, it necessarily follows that the total ow logic is consistent. delta st 00 = X o(l)=s^t dep (l)=t 00^p2P b lp? X d(l)=s^t 0 <t ready (l)t 00^p2P b lp + delta st 0 delta st 00 = X p2p delta spt 00 Zero-Capacity Equality Constraints If a station is at Zero-Capacity (has no available aircraft) at two time-points in the day, then it is possible to deduce that the incoming aircraft are equal to the outgoing aircraft between the two time-points. For example, if a station has no available aircraft at, say, 10am, and a single arrival occurs at 10.30am followed by a single departure at 11am, and there are no other arrivals or departures this interval, the departure at 11am must be carried out by the aircraft arriving at 10.30am. The above example is the simplest case of this form of Zero-Capacity reasoning: 8t 1 ; t 2 2 MT (s) : t 1 < t 2 ^ delta st1 = delta st2 = 0 ) f pt l : o(l) = s ^ t 1 < t dep (l) t 2 g = f pt l : d(l) = s ^ t 1 < t ready (l) t 2 g This type of equality reasoning may be performed at pre-processing stage, or dynamically taking into account run-time assignments. In addition, it may be performed (as described above) independent of plane types, or it may take into the station's capacity with respect to each plane type. The implementation of the pure FD optimizer was enhanced by the simplest form of these constraints i.e. pre-processed Zero-Capacity equality constraints independent of plane type Search Ordering In the pure FD method, search ordering is accomplished by specifying the order in which variables and values are labelled. Search ordering is useful because it may direct the search towards more promising areas.

7 To solve the eet assignment problem, the labelling procedure must label (assign values to) the ight Plane Type variables. 2 There are almost 400 such variables in the BA shorthaul eet assignment, one for each ight. All combinations of the following labelling heuristics were attempted: Variable ordering heuristics Smallest Domain First Most Constrained Variable First Earliest Departure/Arrival First Divide By Station, and Order By Size, Maintenance Capability, etc. Value ordering heuristics Most Available Plane Type First Best Cost Plane Type First 2.3 Results The best combination involved the Maintenance Station First, Ascending Station Size, and Earliest Departure/Arrival First variable ordering heuristics, with no value ordering. This combination reached a rst feasible solution in 329 CPU seconds on a Sun SPARCstation- 20. However, the subsequent optimization process for the above combination was slow (the algorithm showed a 1% improvement at 1 CPU hour and a 3% improvement at 10 CPU hours). 3 The Mixed Integer Programming Method 3.1 Motivation Once the model had been developed and veried the task was to identify an optimization method that produced good quality solutions quickly. The nature of the model, involving a linear optimization function and a set of linear equalities and inequalities, indicated that a mathematical solver capable of solving mixed integer problems might be suitable. At this stage, several other alternative optimization strategies were also candidates for implementation. These were mostly repair algorithms (the method in Section 4 is one example). The solver has an advantage over repair methods in that it is complete; it can prove the global optimality of solutions. 2 Apart from the overnight variables, all other variables are dependent in the sense that they will be instantiated via constraint satisfaction after the complete assignment of the Plane Type variables.

8 Hence the mixed integer programming tool FortMP was chosen for the second method [EH95]. The FD model and a feasible solution were translated for its use. The process of exporting the model and solutions to the FortMP solver was straightforward. The FortMP solver used this model and solutions to quickly obtain the global optimum for this problem; the details of how this was accomplished are detailed in [HEW + 96]. 3.2 Results The FortMP solver obtained the optimum in two hours and 28 minutes. The optimal was 9% better than the user's actual assignment. [HEW + 96] details other results, as well as how the FD feasible solutions were used. The redundant constraints and variables of the pure FD model were incorporated without change. Interestingly, while they resulted in some performance degradation, the degradation was not major (around 10% of the CPU time). 4 The Repair Hybrid 4.1 Motivation A disadvantage of the mixed integer programming method described in Section 3 is that, while it produces the global optimum quickly, its solutions may not bear any relationship to the user's current eet assignment. While knowing the optimal is desirable, a good solution that is close to the current assignment is more useful in practice. Another weakness is that global solutions tend not to be robust. Their real use is as a measure against which other solutions are judged. Hence the need was for a repair algorithm that focussed on local improvements to the user's actual assignment. Furthermore, as shown in this section, the introduction of repair may be computationally benecial to FD propagation. Achieving this benet requires some knowledge of the structure of the problem Analyzing Pure FD Propagation Applied to the eet assignment problem, pure FD implementation is capable of producing a rst feasible solution in reasonable time. However, its performance especially in the optimization phase can be improved. A close examination of the algorithm's search revealed two reasons for the relatively slow performance. Structural Reasons The nature and conguration of the constraints of this problem reduce the eectiveness of constraint propagation during the early stages of search. When a minority of the plane type variables have been labelled, assigning a Plane Type variable representing some ight leg l 0 departing from station s 0 to plane type p 0 often results in the following propagation sequence:

9 { the minimum of a single delta spt variable is incremented to reect the increase in the minimum no. of planes of type p 0 on the ground at the time of departure { the minimum of a single on sp variable is raised by 1 to ensure that a sucient quantity of planes of type p 0 overnight at s 0 { the maxima of up to 60 on sp variables are decremented by 1, to reect the xed total of overnighting aircraft at s 0, and the xed eet size of p 0 { The maxima of several hundred delta spt variables are reduced to reect the decremented maxima of overnight variables Unfortunately, the reduced maxima for delta spt variables translate to useful reductions in the domains of other Plane Type variables in only a small minority cases. This propagation sequence reects the nature of the problem: making a single decision about the plane type of a particular ight may have global implications for the possible quantities of aircraft on the ground before other ight departures (through the eet size and maintenance constraints), but this inferred information does not usually translate into denite information about the possible plane types for other ights until most have already been xed. Once this occurs, however, Finite Domain propagation reductions on the remaining Plane Type variables increase dramatically. Scale A well known deciency of depth rst tree search is that a signicant backtrack overhead is incurred when an early combination of variable labels (close to the search tree root) are inconsistent with all possible later assignments and this inconsistency is not immediately detected. In fact, FD propagation was incorporated into chronological backtracking precisely with a view to addressing this problem; inconsistency is detected earlier using FD propagation. However, in problems where the structure of the problem does not allow FD propagation to detect inconsistencies at an early stage the weakness of depth rst search re-emerges, and becomes more apparent as the size of the problem is increased (eet assignment includes over 400 labelling variables and 7500 secondary variables) Improving Performance Various techniques, such as the introduction of redundant constraints (Section 2.2.1) may be used to hasten FD propagation's detection of inconsistency. Repair is also excellent for this purpose. In fact, knowledge of the structure of the problem may be used together with repair to eectively counter both weaknesses outlined above by (1) localizing FD propagation and (2) improving its eectiveness. 1. Localising Propagation Localising the scope of propagation is possible by labelling those variables in constraints that connect separable portions of the constraint graph [KK88]. In the BA

10 eet assignment the constraint graph may be decomposed into relatively independent subgraphs, corresponding to the stations of the problem, by labelling the overnight variables; these exist in all the constraints linking stations. The stations are still not completely independent though, since labelling or pruning a Plane Type variable has implications for both the departure and arrival stations. 2. Improving Propagation Ecacy The diminished eectiveness of propagation for some problems has been attributed by Minton et. al to lack of informedness, and used as an argument for repair methods [MJPL92]. Tong el al. also report this deciency for global constraints and suggests the use of constrained generators to satisfy them, while applying constraint propagation for the remainder of the problem [TBMV92]. According to Tong et al., a global constraint is one that ranges over a large number of variables without achieving much useful propagation. Constrained generators are procedures which repair an existing solution or modify it such that a global constraint is made redundant. While our constraints do not individually correspond to Tong's denition of a global constraint, the set of constraints on the overnight variables of the eet assignment problem act collectively as a global constraint. 4.2 Incorporating Repair The incorporation of repair for eet assignment depends on a division of the problem variables into repair variables and logical variables. Repair variables start the search with an initial assignment that is propagated onto the remaining logical variables of the problem, reducing their domains. In eect the repair variable assignment reduces search focus to logical sub-problems. Once the logical sub-problem has been searched exhaustively (or a time-out terminates search) logical variable assignments are undone. A reassignment of a subset of the repair variables creates a new logical sub-problem that is searched in a similar manner. In order not to search the same logical sub-problem more than once, those that have already been searched are recorded using nogood labels. Performing a repair labelling on the overnight variables addresses the functional and computational objectives outlined in the previous subsection. First, by starting the search with an overnight allocation equivalent to that of the user's current eet assignment, search nds good solutions close to that region achieving the funcionality required. The overnight variables of the problem are the repair variables that are reassigned to create logical subproblems. Perturbations are made to the user's solution by reassigning subsets of these overnight variables. Second, the problem is decomposed into relatively independent subgraphs corresponding to stations thus limiting the scope of propagation on labelling. This is because the constraints linking stations are on overnight variables, and these are assigned before logical sub-problems are searched. Third, the global constraint part of the problem is handled by eciently by a repair procedure that subsumes the global constraint. If a consistent solution of a constraint is initially given to the xed overnight, eet size and minimum maintenance constraints (Constraints (i),(ii) and (iii) of Section 1.3) operators

11 may be found to perturb this solution without introducing inconsistency; in this way global constraints are subsumed by the repair operator. Fourth, the eectiveness of propagation is enhanced by the tightening of the problem achieved through xing the overnight variables. The domain reductions on plane type and other variables resulting from overnight variable assignments greatly reduce the scope of the subsequent search. The highest level of the repair labelling implementation is outlined here: Nogoods = ;; OvernightLabels = Some Complete Label Set Satisfying Maintenance and Fleet Size Constraints; Repeat OvernightVariables = OvernightLabels; LabelPlaneTypeVariables(TimeOut, NewCost, BoolSuccess) if(boolsuccess) then AddConstraint(CostVariable NewCost) else Nogoods = Nogoods [ OvernightLabels; Repeat NewOvernightLabels = RandomSwapTwoPlaneLocations(OvernightLabels); Until NewOvernightLabels =2 Nogoods; OvernightLabels = NewOvernightLabels; Until False; LabelPlaneTypeVariables(TimeOut, NewCost, BoolSuccess) is a procedure that sets BoolSuccess to be True only when NewCost is a new cost that has been found through labelling the Plane Type variables in CPU time less than Timeout; otherwise the procedure returns before a xed time Timeout with BoolSuccess set to False. AddConstraint(Constraint) is a procedure that adds a constraint to the FD constraint store. Finally, RandomSwapTwo- PlaneLocations(OvernightLabels) is a repair procedure that selects a random pair of overnight variables from dierent stations, decrements one and increments the other thus eecting a swap between of two planes at dierent stations and of dierent plane type. Unlike many pure Logic Programming environments ECL i PS e makes the incorporation of repair easy through its support of destructive variable assignment, meta-variable structures, and through a high degree of propagation control for individual constraints. 4.3 Results Following the introduction of the repair component, the performance of FD propagation improved dramatically. A rst solution is reached in under 1 minute, with a 4% improvement over the rst feasible solution achieved in 10 minutes and a 6% improvement after 1 CPU hour. Heuristics such as Best-Cost-Value-First and Smallest-Domain-Variable-First achieved an increase in performance due to the tightening of the constraint network resulting from overnight variable labelling. The solutions produced at the start of the search

12 were close to the actual eet assignment, which meant they were of practical use. 5 Conclusion Generally speaking, a conict of interests exists in systems which generalise over several alternative solution techniques: exibility in switching between them versus specializing the problem description for ecient execution by any single one. By providing a separation and a high quality interface between problem descriptions and the techniques used for solving them, the ECL i PS e environment achieved exibility in this instance. It allowed alternative models to be quickly tested for correctness and eciency, and facilitated the application of several alternative algorithms to a chosen model. Moreover, the eet assignment problem showed that there is in fact much common ground among solution methods: Many of the solution method parameters were similar (decision variables, search ordering heuristics, etc.). For example, the overnight and (Boolean) plane type variables and emerged as critical variables that had to be given priority across all three methods, either through early labelling in the rst and third methods, or through explicit integer declarations in the second. Model specialization did not signicantly degrade performance of other methods. The FortMP solver was not signicantly disadvantaged by the introduction of redundant constraints, for example. Finally, the benets resulting from a exible platform outweigh any minor losses in eciency; we believe it is initially far more important to quickly and simply apply a range of possibly hybrid solution methods to a new problem than to focus on specializing any single one. The BA shorthaul eet assignment problem solved here is in fact an approximation of the \real problem" tailored in the past for easy representation and solution by a mathematical solver. This work has convinced the user to invest in the development of an ECL i PS e -based prototype to improve their more detailed day to day eet assignment process, using a repair FD propagation hybrid approach akin to the third method described in this paper. In conclusion: the process of rapidly constructing and testing tailored algorithms is supported in the current ECL i PS e environment. Future versions of ECL i PS e which further facilitate the tasks of solution exploration and integration are currently being developed at IC-Parc. References [Dav87] E. Davis. Constraint propagation with interval labels. Articial Intelligence, 32:281{331, 1987.

13 [ECR95] ECRC. ECLiPSe 3.5. Technical report, ECRC, [EH95] [HE80] E.F.D Ellison and M. Hajian. Fortmp, user guide and reference manual. Technical report, Nag Ltd, Oxford, UK, R.M. Haralick and G.L. Elliot. Increasing tree search eciency for constraint satisfaction problems. Articial Intelligence, 14:263{314, October [HEW + 96] M. Hajian, H. El Sakkout, M. Wallace, J. Lever, and E.B. Richards. Towards a closer integration of nite domain propagation and simpled-based algorithms. In Proceedings of the Fourth International Symposium on Articial Intelligence and Mathematics, Florida, January [JL87] J. Jaar and J.-L. Lassez. Constraint logic programming. In Proceedings of the Fourteenth ACM Symposium on Principles of Programming Languages (POPL'87), Munich, FRG, January [KK88] L. Kanal and V. Kumar. Search in Articial Intelligence. Springer-Verlag, [Kow79] R. A. Kowalski. Logic for Problem Solving. North-Holland, [Mac77] [MJPL92] [SS80] A.K. Mackworth. Consistency in networks of relations. Articial Intelligence, 8(1):99{118, Steven Minton, Mark D. Johnston, Andrew B. Philips, and Philip Laird. Minimizing conicts: a heuristic repair method for constraint satisfaction and scheduling problems. Articial Intelligence, 58:161{205, G.J. Sussman and G.L. Steele. CONSTRAINTS: A language for expressing almost-hierarchical descriptions. Articial Intelligence, 14(1):1{39, January [TBMV92] Chris Tong, Wesley Braudaway, Sunil Mohan, and Kerstin Voigt. Reformulating constraints for compilability and eciency. AAAI '92, Workshop on Problem Reformulation and Change of Presentation, [Van89] P. Van Hentenryck. Constraint Satisfaction in Logic Programming. Logic Programming Series. The MIT Press, 1989.

Two Problems - Two Solutions: One System - ECLiPSe. Mark Wallace and Andre Veron. April 1993

Two Problems - Two Solutions: One System - ECLiPSe. Mark Wallace and Andre Veron. April 1993 Two Problems - Two Solutions: One System - ECLiPSe Mark Wallace and Andre Veron April 1993 1 Introduction The constraint logic programming system ECL i PS e [4] is the successor to the CHIP system [1].

More information

Modelling Combinatorial Problems for CLP(FD+R) Henk Vandecasteele. Department of Computer Science, K. U. Leuven

Modelling Combinatorial Problems for CLP(FD+R) Henk Vandecasteele. Department of Computer Science, K. U. Leuven Modelling Combinatorial Problems for CLP(FD+R) Henk Vandecasteele Department of Computer Science, K. U. Leuven Celestijnenlaan 200A, B-3001 Heverlee, Belgium henk.vandecasteele@cs.kuleuven.ac.be Robert

More information

2WKHUZLVH 6HOHFWFRQIOLFW& LL 6HOHFWFRQVWUDLQW7& WKDWZRXOGOHDGWRUHSDLU& LLL &UHDWHDFKRLFHSRLQWFRUUHVSRQGLQJD GLVMXQFWLRQRI7& DQGLWVQHJDWLRQ*RWR

2WKHUZLVH 6HOHFWFRQIOLFW& LL 6HOHFWFRQVWUDLQW7& WKDWZRXOGOHDGWRUHSDLU& LLL &UHDWHDFKRLFHSRLQWFRUUHVSRQGLQJD GLVMXQFWLRQRI7& DQGLWVQHJDWLRQ*RWR Local Probing for Resource Constrained Scheduling Olli Kamarainen, Hani El Sakkout, Jonathan Lever IC-Parc Imperial College of Science, Technology and Medicine, London SW7 2AZ, UK {ok1,hhe,jml}@icparc.ic.ac.uk

More information

Unimodular probing for minimal perturbance. in dynamic resource feasibility problems. Hani El Sakkout, Thomas Richards, Mark G.

Unimodular probing for minimal perturbance. in dynamic resource feasibility problems. Hani El Sakkout, Thomas Richards, Mark G. Unimodular probing for minimal perturbance in dynamic resource feasibility problems Hani El Sakkout, Thomas Richards, Mark G. Wallace IC-Parc, Imperial College, London SW7 2AZ, United Kingdom. Email: fhhe,mgwg

More information

2 Keywords Backtracking Algorithms, Constraint Satisfaction Problem, Distributed Articial Intelligence, Iterative Improvement Algorithm, Multiagent Sy

2 Keywords Backtracking Algorithms, Constraint Satisfaction Problem, Distributed Articial Intelligence, Iterative Improvement Algorithm, Multiagent Sy 1 The Distributed Constraint Satisfaction Problem: Formalization and Algorithms IEEE Trans. on Knowledge and DATA Engineering, vol.10, No.5 September 1998 Makoto Yokoo, Edmund H. Durfee, Toru Ishida, and

More information

Kalev Kask and Rina Dechter. Department of Information and Computer Science. University of California, Irvine, CA

Kalev Kask and Rina Dechter. Department of Information and Computer Science. University of California, Irvine, CA GSAT and Local Consistency 3 Kalev Kask and Rina Dechter Department of Information and Computer Science University of California, Irvine, CA 92717-3425 fkkask,dechterg@ics.uci.edu Abstract It has been

More information

Andrew Davenport and Edward Tsang. fdaveat,edwardgessex.ac.uk. mostly soluble problems and regions of overconstrained, mostly insoluble problems as

Andrew Davenport and Edward Tsang. fdaveat,edwardgessex.ac.uk. mostly soluble problems and regions of overconstrained, mostly insoluble problems as An empirical investigation into the exceptionally hard problems Andrew Davenport and Edward Tsang Department of Computer Science, University of Essex, Colchester, Essex CO SQ, United Kingdom. fdaveat,edwardgessex.ac.uk

More information

GSAT and Local Consistency

GSAT and Local Consistency GSAT and Local Consistency Kalev Kask Computer Science Department University of California at Irvine Irvine, CA 92717 USA Rina Dechter Computer Science Department University of California at Irvine Irvine,

More information

A generic framework for solving CSPs integrating decomposition methods

A generic framework for solving CSPs integrating decomposition methods A generic framework for solving CSPs integrating decomposition methods L. Blet 1,3, S. N. Ndiaye 1,2, and C. Solnon 1,3 1 Université de Lyon - LIRIS 2 Université Lyon 1, LIRIS, UMR5205, F-69622 France

More information

Rina Dechter. Irvine, California, USA A constraint satisfaction problem (csp) dened over a constraint network

Rina Dechter. Irvine, California, USA A constraint satisfaction problem (csp) dened over a constraint network Constraint Satisfaction Rina Dechter Department of Computer and Information Science University of California, Irvine Irvine, California, USA 92717 dechter@@ics.uci.edu A constraint satisfaction problem

More information

Some Applications of Graph Bandwidth to Constraint Satisfaction Problems

Some Applications of Graph Bandwidth to Constraint Satisfaction Problems Some Applications of Graph Bandwidth to Constraint Satisfaction Problems Ramin Zabih Computer Science Department Stanford University Stanford, California 94305 Abstract Bandwidth is a fundamental concept

More information

These advantages are exemplied by a scheduling workbench called Oz Scheduler [13]. This workbench makes use of dierent OR algorithms for propagation a

These advantages are exemplied by a scheduling workbench called Oz Scheduler [13]. This workbench makes use of dierent OR algorithms for propagation a Constraint-Based Scheduling in Oz Jorg Wurtz, DFKI GmbH, Saarbrucken Abstract It is discussed, how scheduling problems can be solved in the concurrent constraint programming language Oz. Oz is the rst

More information

of m clauses, each containing the disjunction of boolean variables from a nite set V = fv 1 ; : : : ; vng of size n [8]. Each variable occurrence with

of m clauses, each containing the disjunction of boolean variables from a nite set V = fv 1 ; : : : ; vng of size n [8]. Each variable occurrence with A Hybridised 3-SAT Algorithm Andrew Slater Automated Reasoning Project, Computer Sciences Laboratory, RSISE, Australian National University, 0200, Canberra Andrew.Slater@anu.edu.au April 9, 1999 1 Introduction

More information

2 The Service Provision Problem The formulation given here can also be found in Tomasgard et al. [6]. That paper also details the background of the mo

2 The Service Provision Problem The formulation given here can also be found in Tomasgard et al. [6]. That paper also details the background of the mo Two-Stage Service Provision by Branch and Bound Shane Dye Department ofmanagement University of Canterbury Christchurch, New Zealand s.dye@mang.canterbury.ac.nz Asgeir Tomasgard SINTEF, Trondheim, Norway

More information

Steering. Stream. User Interface. Stream. Manager. Interaction Managers. Snapshot. Stream

Steering. Stream. User Interface. Stream. Manager. Interaction Managers. Snapshot. Stream Agent Roles in Snapshot Assembly Delbert Hart Dept. of Computer Science Washington University in St. Louis St. Louis, MO 63130 hart@cs.wustl.edu Eileen Kraemer Dept. of Computer Science University of Georgia

More information

Constraint Optimisation Problems. Constraint Optimisation. Cost Networks. Branch and Bound. Dynamic Programming

Constraint Optimisation Problems. Constraint Optimisation. Cost Networks. Branch and Bound. Dynamic Programming Summary Network Search Bucket Elimination Soft Soft s express preferences over variable assignments Preferences give dierent values over variable assignment A student can follow only one class at a time

More information

size, runs an existing induction algorithm on the rst subset to obtain a rst set of rules, and then processes each of the remaining data subsets at a

size, runs an existing induction algorithm on the rst subset to obtain a rst set of rules, and then processes each of the remaining data subsets at a Multi-Layer Incremental Induction Xindong Wu and William H.W. Lo School of Computer Science and Software Ebgineering Monash University 900 Dandenong Road Melbourne, VIC 3145, Australia Email: xindong@computer.org

More information

Network. Department of Statistics. University of California, Berkeley. January, Abstract

Network. Department of Statistics. University of California, Berkeley. January, Abstract Parallelizing CART Using a Workstation Network Phil Spector Leo Breiman Department of Statistics University of California, Berkeley January, 1995 Abstract The CART (Classication and Regression Trees) program,

More information

Solve the Data Flow Problem

Solve the Data Flow Problem Gaining Condence in Distributed Systems Gleb Naumovich, Lori A. Clarke, and Leon J. Osterweil University of Massachusetts, Amherst Computer Science Department University of Massachusetts Amherst, Massachusetts

More information

Constraint Satisfaction Problems. Chapter 6

Constraint Satisfaction Problems. Chapter 6 Constraint Satisfaction Problems Chapter 6 Constraint Satisfaction Problems A constraint satisfaction problem consists of three components, X, D, and C: X is a set of variables, {X 1,..., X n }. D is a

More information

A Boolean Expression. Reachability Analysis or Bisimulation. Equation Solver. Boolean. equations.

A Boolean Expression. Reachability Analysis or Bisimulation. Equation Solver. Boolean. equations. A Framework for Embedded Real-time System Design? Jin-Young Choi 1, Hee-Hwan Kwak 2, and Insup Lee 2 1 Department of Computer Science and Engineering, Korea Univerity choi@formal.korea.ac.kr 2 Department

More information

Title of Publication: Encyclopedia of Cognitive Science. Article title: Constraint Satisfaction. Article code: 26. Authors: Rina Dechter

Title of Publication: Encyclopedia of Cognitive Science. Article title: Constraint Satisfaction. Article code: 26. Authors: Rina Dechter Title of Publication: Encyclopedia of Cognitive Science. Article title: Constraint Satisfaction. Article code: 26. Authors: Rina Dechter Department of Computer and Information Science University of California,

More information

A. Atamturk. G.L. Nemhauser. M.W.P. Savelsbergh. Georgia Institute of Technology. School of Industrial and Systems Engineering.

A. Atamturk. G.L. Nemhauser. M.W.P. Savelsbergh. Georgia Institute of Technology. School of Industrial and Systems Engineering. A Combined Lagrangian, Linear Programming and Implication Heuristic for Large-Scale Set Partitioning Problems 1 A. Atamturk G.L. Nemhauser M.W.P. Savelsbergh Georgia Institute of Technology School of Industrial

More information

General Methods and Search Algorithms

General Methods and Search Algorithms DM811 HEURISTICS AND LOCAL SEARCH ALGORITHMS FOR COMBINATORIAL OPTIMZATION Lecture 3 General Methods and Search Algorithms Marco Chiarandini 2 Methods and Algorithms A Method is a general framework for

More information

space. We will apply the idea of enforcing local consistency to GSAT with the hope that its performance can

space. We will apply the idea of enforcing local consistency to GSAT with the hope that its performance can GSAT and Local Consistency 3 Kalev Kask Computer Science Department University of California at Irvine Irvine, CA 92717 USA Rina Dechter Computer Science Department University of California at Irvine Irvine,

More information

Path Consistency Revisited. Moninder Singh. University of Pennsylvania. Philadelphia, PA

Path Consistency Revisited. Moninder Singh. University of Pennsylvania. Philadelphia, PA Proceedings of the 7th IEEE International Conference on Tools with Articial Intelligence, 318-325, 1995. c Institute of Electrical and Electronics Engineers, Inc. (IEEE). Path Consistency Revisited Moninder

More information

Brian Drabble, Je Dalton and Austin Tate

Brian Drabble, Je Dalton and Austin Tate Technical Report O-Plan Tasking Specication Brian Drabble, Je Dalton and Austin Tate Approved for public release; distribution is unlimited Articial Intelligence Applications Institute University of Edinburgh

More information

Validating Plans with Durative Actions via Integrating Boolean and Numerical Constraints

Validating Plans with Durative Actions via Integrating Boolean and Numerical Constraints Validating Plans with Durative Actions via Integrating Boolean and Numerical Constraints Roman Barták Charles University in Prague, Faculty of Mathematics and Physics Institute for Theoretical Computer

More information

Exploring A Two-Solver Architecture for Clause Learning CSP Solvers. Ozan Erdem

Exploring A Two-Solver Architecture for Clause Learning CSP Solvers. Ozan Erdem Exploring A Two-Solver Architecture for Clause Learning CSP Solvers by Ozan Erdem A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Graduate Department of Computer

More information

Richard E. Korf. June 27, Abstract. divide them into two subsets, so that the sum of the numbers in

Richard E. Korf. June 27, Abstract. divide them into two subsets, so that the sum of the numbers in A Complete Anytime Algorithm for Number Partitioning Richard E. Korf Computer Science Department University of California, Los Angeles Los Angeles, Ca. 90095 korf@cs.ucla.edu June 27, 1997 Abstract Given

More information

Constraint Logic Programming and Integer Programming approaches and their collaboration in solving an assignment scheduling problem

Constraint Logic Programming and Integer Programming approaches and their collaboration in solving an assignment scheduling problem Constraint Logic Programming and Integer Programming approaches and their collaboration in solving an assignment scheduling problem Ken Darby-Dowman (1) James Little (1) Gautam Mitra (1) Marco Zaffalon

More information

Lecture 6: Constraint Satisfaction Problems (CSPs)

Lecture 6: Constraint Satisfaction Problems (CSPs) Lecture 6: Constraint Satisfaction Problems (CSPs) CS 580 (001) - Spring 2018 Amarda Shehu Department of Computer Science George Mason University, Fairfax, VA, USA February 28, 2018 Amarda Shehu (580)

More information

Consistent Logical Checkpointing. Nitin H. Vaidya. Texas A&M University. Phone: Fax:

Consistent Logical Checkpointing. Nitin H. Vaidya. Texas A&M University. Phone: Fax: Consistent Logical Checkpointing Nitin H. Vaidya Department of Computer Science Texas A&M University College Station, TX 77843-3112 hone: 409-845-0512 Fax: 409-847-8578 E-mail: vaidya@cs.tamu.edu Technical

More information

A Re-examination of Limited Discrepancy Search

A Re-examination of Limited Discrepancy Search A Re-examination of Limited Discrepancy Search W. Ken Jackson, Morten Irgens, and William S. Havens Intelligent Systems Lab, Centre for Systems Science Simon Fraser University Burnaby, B.C., CANADA V5A

More information

Constraint Satisfaction Problems

Constraint Satisfaction Problems Constraint Satisfaction Problems CE417: Introduction to Artificial Intelligence Sharif University of Technology Spring 2013 Soleymani Course material: Artificial Intelligence: A Modern Approach, 3 rd Edition,

More information

Algorithmic "imperative" language

Algorithmic imperative language Algorithmic "imperative" language Undergraduate years Epita November 2014 The aim of this document is to introduce breiy the "imperative algorithmic" language used in the courses and tutorials during the

More information

Local consistency for ternary numeric constraints. Boi Faltings and Esther Gelle. Articial Intelligence Laboratory (LIA)

Local consistency for ternary numeric constraints. Boi Faltings and Esther Gelle. Articial Intelligence Laboratory (LIA) Local consistency for ternary numeric constraints Boi Faltings and Esther Gelle Articial Intelligence Laboratory (LIA) Swiss Federal Institute of Technology (EPFL) IN-Ecublens, 1015 Lausanne, Switzerland

More information

In search of the best constraint satisfaction search 3. Daniel Frost and Rina Dechter.

In search of the best constraint satisfaction search 3. Daniel Frost and Rina Dechter. In search of the best constraint satisfaction search 3 Daniel Frost and Rina Dechter Dept. of Information and Computer Science University of California, Irvine, CA 92717 fdfrost,dechterg@ics.uci.edu Abstract

More information

Submitted for TAU97 Abstract Many attempts have been made to combine some form of retiming with combinational

Submitted for TAU97 Abstract Many attempts have been made to combine some form of retiming with combinational Experiments in the Iterative Application of Resynthesis and Retiming Soha Hassoun and Carl Ebeling Department of Computer Science and Engineering University ofwashington, Seattle, WA fsoha,ebelingg@cs.washington.edu

More information

Integer Programming ISE 418. Lecture 7. Dr. Ted Ralphs

Integer Programming ISE 418. Lecture 7. Dr. Ted Ralphs Integer Programming ISE 418 Lecture 7 Dr. Ted Ralphs ISE 418 Lecture 7 1 Reading for This Lecture Nemhauser and Wolsey Sections II.3.1, II.3.6, II.4.1, II.4.2, II.5.4 Wolsey Chapter 7 CCZ Chapter 1 Constraint

More information

Algebraic Properties of CSP Model Operators? Y.C. Law and J.H.M. Lee. The Chinese University of Hong Kong.

Algebraic Properties of CSP Model Operators? Y.C. Law and J.H.M. Lee. The Chinese University of Hong Kong. Algebraic Properties of CSP Model Operators? Y.C. Law and J.H.M. Lee Department of Computer Science and Engineering The Chinese University of Hong Kong Shatin, N.T., Hong Kong SAR, China fyclaw,jleeg@cse.cuhk.edu.hk

More information

Constraint Satisfaction

Constraint Satisfaction Constraint Satisfaction Philipp Koehn 1 October 2015 Outline 1 Constraint satisfaction problems (CSP) examples Backtracking search for CSPs Problem structure and problem decomposition Local search for

More information

A taxonomy of race. D. P. Helmbold, C. E. McDowell. September 28, University of California, Santa Cruz. Santa Cruz, CA

A taxonomy of race. D. P. Helmbold, C. E. McDowell. September 28, University of California, Santa Cruz. Santa Cruz, CA A taxonomy of race conditions. D. P. Helmbold, C. E. McDowell UCSC-CRL-94-34 September 28, 1994 Board of Studies in Computer and Information Sciences University of California, Santa Cruz Santa Cruz, CA

More information

Course Summary! What have we learned and what are we expected to know?

Course Summary! What have we learned and what are we expected to know? Course Summary! What have we learned and what are we expected to know? Overview! Introduction Modelling in MiniZinc Finite Domain Constraint Solving Search Linear Programming and Network Flow Mixed Integer

More information

On Computing Minimum Size Prime Implicants

On Computing Minimum Size Prime Implicants On Computing Minimum Size Prime Implicants João P. Marques Silva Cadence European Laboratories / IST-INESC Lisbon, Portugal jpms@inesc.pt Abstract In this paper we describe a new model and algorithm for

More information

A New Algorithm for Singleton Arc Consistency

A New Algorithm for Singleton Arc Consistency A New Algorithm for Singleton Arc Consistency Roman Barták, Radek Erben Charles University, Institute for Theoretical Computer Science Malostranské nám. 2/25, 118 Praha 1, Czech Republic bartak@kti.mff.cuni.cz,

More information

University of Leeds SCHOOL OF COMPUTER STUDIES RESEARCH REPORT SERIES. Report 99.05

University of Leeds SCHOOL OF COMPUTER STUDIES RESEARCH REPORT SERIES. Report 99.05 University of eeds SCHOO OF COMPUTER STUDIES RESEARCH REPORT SERIES Report 99.05 Forming Bus Driver Schedules using Constraint Programming by Suniel D Curtis, Barbara M Smith, Anthony Wren March 1999 Abstract

More information

Constraint (Logic) Programming

Constraint (Logic) Programming Constraint (Logic) Programming Roman Barták Faculty of Mathematics and Physics, Charles University in Prague, Czech Republic bartak@ktiml.mff.cuni.cz Sudoku Combinatorial puzzle, whose goal is to enter

More information

Journal of Global Optimization, 10, 1{40 (1997) A Discrete Lagrangian-Based Global-Search. Method for Solving Satisability Problems *

Journal of Global Optimization, 10, 1{40 (1997) A Discrete Lagrangian-Based Global-Search. Method for Solving Satisability Problems * Journal of Global Optimization, 10, 1{40 (1997) c 1997 Kluwer Academic Publishers, Boston. Manufactured in The Netherlands. A Discrete Lagrangian-Based Global-Search Method for Solving Satisability Problems

More information

DRAFT for FINAL VERSION. Accepted for CACSD'97, Gent, Belgium, April 1997 IMPLEMENTATION ASPECTS OF THE PLC STANDARD IEC

DRAFT for FINAL VERSION. Accepted for CACSD'97, Gent, Belgium, April 1997 IMPLEMENTATION ASPECTS OF THE PLC STANDARD IEC DRAFT for FINAL VERSION. Accepted for CACSD'97, Gent, Belgium, 28-3 April 1997 IMPLEMENTATION ASPECTS OF THE PLC STANDARD IEC 1131-3 Martin hman Stefan Johansson Karl-Erik rzen Department of Automatic

More information

Creating Meaningful Training Data for Dicult Job Shop Scheduling Instances for Ordinal Regression

Creating Meaningful Training Data for Dicult Job Shop Scheduling Instances for Ordinal Regression Creating Meaningful Training Data for Dicult Job Shop Scheduling Instances for Ordinal Regression Helga Ingimundardóttir University of Iceland March 28 th, 2012 Outline Introduction Job Shop Scheduling

More information

Constraint-Directed Backtracking Algorithm. Wanlin Pang. Scott D. Goodwin. May 1996

Constraint-Directed Backtracking Algorithm. Wanlin Pang. Scott D. Goodwin. May 1996 Constraint-Directed Backtracking Algorithm for Constraint-Satisfaction Problems Wanlin Pang Scott D. Goodwin Technical Report CS-96-05 May 1996 cwanlin Pang and Scott D. Goodwin Department of Computer

More information

Rajiv Chopra, Rohini Srihari and Anthony Ralston. labeling problem. However, these are dicult to

Rajiv Chopra, Rohini Srihari and Anthony Ralston. labeling problem. However, these are dicult to Expensive Constraints and HyperArc-Consistency Rajiv Chopra, Rohini Srihari and Anthony Ralston Dept. of Computer Science, SUNY at Bualo 224 Bell Hall, Amherst NY 14214 frchopra,rohini,ralstong@cs.buffalo.edu

More information

Neuro-Remodeling via Backpropagation of Utility. ABSTRACT Backpropagation of utility is one of the many methods for neuro-control.

Neuro-Remodeling via Backpropagation of Utility. ABSTRACT Backpropagation of utility is one of the many methods for neuro-control. Neuro-Remodeling via Backpropagation of Utility K. Wendy Tang and Girish Pingle 1 Department of Electrical Engineering SUNY at Stony Brook, Stony Brook, NY 11794-2350. ABSTRACT Backpropagation of utility

More information

lpsymphony - Integer Linear Programming in R

lpsymphony - Integer Linear Programming in R lpsymphony - Integer Linear Programming in R Vladislav Kim October 30, 2017 Contents 1 Introduction 2 2 lpsymphony: Quick Start 2 3 Integer Linear Programming 5 31 Equivalent and Dual Formulations 5 32

More information

Constraint Satisfaction Problems. slides from: Padhraic Smyth, Bryan Low, S. Russell and P. Norvig, Jean-Claude Latombe

Constraint Satisfaction Problems. slides from: Padhraic Smyth, Bryan Low, S. Russell and P. Norvig, Jean-Claude Latombe Constraint Satisfaction Problems slides from: Padhraic Smyth, Bryan Low, S. Russell and P. Norvig, Jean-Claude Latombe Standard search problems: State is a black box : arbitrary data structure Goal test

More information

ALGORITHM SYSTEMS FOR COMBINATORIAL OPTIMIZATION: HIERARCHICAL MULTISTAGE FRAMEWORK

ALGORITHM SYSTEMS FOR COMBINATORIAL OPTIMIZATION: HIERARCHICAL MULTISTAGE FRAMEWORK ALGORITHM SYSTEMS FOR COMBINATORIAL OPTIMIZATION: HIERARCHICAL MULTISTAGE FRAMEWORK Dr. Mark Sh. Levin, The Research Inst., The College Judea & Samaria, Ariel, Israel Introduction In recent decades, signicance

More information

Connecting special ordered inequalities and transformation and reformulation technique in multiple choice programming

Connecting special ordered inequalities and transformation and reformulation technique in multiple choice programming Computers & Operations Research 29 (2002) 1441}1446 Short communication Connecting special ordered inequalities and transformation and reformulation technique in multiple choice programming Edward Yu-Hsien

More information

Constraint Satisfaction Problems Chapter 3, Section 7 and Chapter 4, Section 4.4 AIMA Slides cstuart Russell and Peter Norvig, 1998 Chapter 3, Section

Constraint Satisfaction Problems Chapter 3, Section 7 and Chapter 4, Section 4.4 AIMA Slides cstuart Russell and Peter Norvig, 1998 Chapter 3, Section Constraint Satisfaction Problems Chapter 3, Section 7 and Chapter 4, Section 4.4 AIMA Slides cstuart Russell and Peter Norvig, 1998 Chapter 3, Section 7 and Chapter 4, Section 4.4 1 Outline } CSP examples

More information

A CSP Search Algorithm with Reduced Branching Factor

A CSP Search Algorithm with Reduced Branching Factor A CSP Search Algorithm with Reduced Branching Factor Igor Razgon and Amnon Meisels Department of Computer Science, Ben-Gurion University of the Negev, Beer-Sheva, 84-105, Israel {irazgon,am}@cs.bgu.ac.il

More information

Parallel Branch & Bound

Parallel Branch & Bound Parallel Branch & Bound Bernard Gendron Université de Montréal gendron@iro.umontreal.ca Outline Mixed integer programming (MIP) and branch & bound (B&B) Linear programming (LP) based B&B Relaxation and

More information

Constraint-Based Scheduling: An Introduction for Newcomers

Constraint-Based Scheduling: An Introduction for Newcomers Constraint-Based Scheduling: An Introduction for Newcomers Roman Barták * Charles University in Prague, Faculty of Mathematics and Physics Malostranské námestí 2/25, 118 00, Praha 1, Czech Republic bartak@kti.mff.cuni.cz

More information

is a kind of generalization of AC-3 to non-binary constraints. As AC-3, that algorithm has a bad worst-case time complexity (O(er 2 d r+1 ), with e th

is a kind of generalization of AC-3 to non-binary constraints. As AC-3, that algorithm has a bad worst-case time complexity (O(er 2 d r+1 ), with e th Arc consistency for general constraint networks: preliminary results Christian Bessiere LIRMM{CNRS (UMR 5506) 161 rue Ada 34392 Montpellier cedex 5, France Email: bessiere@lirmm.fr Jean-Charles Regin ILOG

More information

Mathematical Programming Formulations, Constraint Programming

Mathematical Programming Formulations, Constraint Programming Outline DM87 SCHEDULING, TIMETABLING AND ROUTING Lecture 3 Mathematical Programming Formulations, Constraint Programming 1. Special Purpose Algorithms 2. Constraint Programming Marco Chiarandini DM87 Scheduling,

More information

16 Greedy Algorithms

16 Greedy Algorithms 16 Greedy Algorithms Optimization algorithms typically go through a sequence of steps, with a set of choices at each For many optimization problems, using dynamic programming to determine the best choices

More information

Topics in Constraint Programming: Exercises

Topics in Constraint Programming: Exercises Topics in Constraint Programming: Exercises K. Kuchcinski, J. Ma luszyński, U. Nilsson October, 999 README.FIRST The objective of the attached exercises is to give some practical experience with: the constraint

More information

perform. If more storage is required, more can be added without having to modify the processor (provided that the extra memory is still addressable).

perform. If more storage is required, more can be added without having to modify the processor (provided that the extra memory is still addressable). How to Make Zuse's Z3 a Universal Computer Raul Rojas January 14, 1998 Abstract The computing machine Z3, built by Konrad Zuse between 1938 and 1941, could only execute xed sequences of oating-point arithmetical

More information

Solving Large Aircraft Landing Problems on Multiple Runways by Applying a Constraint Programming Approach

Solving Large Aircraft Landing Problems on Multiple Runways by Applying a Constraint Programming Approach Solving Large Aircraft Landing Problems on Multiple Runways by Applying a Constraint Programming Approach Amir Salehipour School of Mathematical and Physical Sciences, The University of Newcastle, Australia

More information

In the International Journal of Parallel Programming, vol.28, no. 1, Enhanced Co-Scheduling: A Software Pipelining Method

In the International Journal of Parallel Programming, vol.28, no. 1, Enhanced Co-Scheduling: A Software Pipelining Method In the International Journal of Parallel Programming, vol.28, no. 1, pages 1{46, Feb. 2000. Enhanced Co-Scheduling: A Software Pipelining Method using Modulo-Scheduled Pipeline Theory R. Govindarajan N.S.S.

More information

Research on outlier intrusion detection technologybased on data mining

Research on outlier intrusion detection technologybased on data mining Acta Technica 62 (2017), No. 4A, 635640 c 2017 Institute of Thermomechanics CAS, v.v.i. Research on outlier intrusion detection technologybased on data mining Liang zhu 1, 2 Abstract. With the rapid development

More information

A tabu search based memetic algorithm for the max-mean dispersion problem

A tabu search based memetic algorithm for the max-mean dispersion problem A tabu search based memetic algorithm for the max-mean dispersion problem Xiangjing Lai a and Jin-Kao Hao a,b, a LERIA, Université d'angers, 2 Bd Lavoisier, 49045 Angers, France b Institut Universitaire

More information

Constraint Satisfaction Problems Part 2

Constraint Satisfaction Problems Part 2 Constraint Satisfaction Problems Part 2 Deepak Kumar October 2017 CSP Formulation (as a special case of search) State is defined by n variables x 1, x 2,, x n Variables can take on values from a domain

More information

Constraint Programming

Constraint Programming Constraint In Pursuit of The Holly Grail Roman Barták Charles University in Prague Constraint programming represents one of the closest approaches computer science has yet made to the Holy Grail of programming:

More information

Constraint Satisfaction Problems

Constraint Satisfaction Problems Constraint Satisfaction Problems Chapter 5 Chapter 5 1 Outline CSP examples Backtracking search for CSPs Problem structure and problem decomposition Local search for CSPs Chapter 5 2 Constraint satisfaction

More information

Hybrid Constraint Solvers

Hybrid Constraint Solvers Hybrid Constraint Solvers - An overview Why Hybrid Solvers CP and SAT: Lazy Clause Generation CP and LP: Reification of Linear Constraints Conclusions 9 November 2011 Pedro Barahona - EPCL - Hybrid Solvers

More information

SCHEDULING REAL-TIME MESSAGES IN PACKET-SWITCHED NETWORKS IAN RAMSAY PHILP. B.S., University of North Carolina at Chapel Hill, 1988

SCHEDULING REAL-TIME MESSAGES IN PACKET-SWITCHED NETWORKS IAN RAMSAY PHILP. B.S., University of North Carolina at Chapel Hill, 1988 SCHEDULING REAL-TIME MESSAGES IN PACKET-SWITCHED NETWORKS BY IAN RAMSAY PHILP B.S., University of North Carolina at Chapel Hill, 1988 M.S., University of Florida, 1990 THESIS Submitted in partial fulllment

More information

A Bintree Representation of Generalized Binary. Digital Images

A Bintree Representation of Generalized Binary. Digital Images A intree Representation of Generalized inary Digital mages Hanspeter ieri gor Metz 1 inary Digital mages and Hyperimages A d-dimensional binary digital image can most easily be modelled by a d-dimensional

More information

A Hybrid Algorithm for Compiling Equality. Constraints and its Implementation. Department of Computer and Information Science

A Hybrid Algorithm for Compiling Equality. Constraints and its Implementation. Department of Computer and Information Science URL: http://www.elsevier.nl/locate/entcs/volume30.html 9 pages A Hybrid Algorithm for Compiling Equality Constraints and its Implementation Neng-Fa Zhou Department of Computer and Information Science Brooklyn

More information

i 1 CONSTRAINT SATISFACTION METHODS FOR GENERATING VALID CUTS J. N. Hooker Graduate School of Industrial Administration Carnegie Mellon University Pittsburgh, PA 15213 USA http://www.gsia.cmu.edu/afs/andrew/gsia/jh38/jnh.html

More information

Conflict-based Statistics

Conflict-based Statistics Conflict-based Statistics Tomáš Müller 1, Roman Barták 1 and Hana Rudová 2 1 Faculty of Mathematics and Physics, Charles University Malostranské nám. 2/25, Prague, Czech Republic {muller bartak}@ktiml.mff.cuni.cz

More information

Unrestricted Nogood Recording in CSP search

Unrestricted Nogood Recording in CSP search Unrestricted Nogood Recording in CSP search George Katsirelos and Fahiem Bacchus Department of Computer Science, University Of Toronto, Toronto, Ontario, Canada [gkatsi,fbacchus]@cs.toronto.edu Abstract.

More information

Greedy Algorithms. T. M. Murali. January 28, Interval Scheduling Interval Partitioning Minimising Lateness

Greedy Algorithms. T. M. Murali. January 28, Interval Scheduling Interval Partitioning Minimising Lateness Greedy Algorithms T. M. Murali January 28, 2008 Algorithm Design Start discussion of dierent ways of designing algorithms. Greedy algorithms, divide and conquer, dynamic programming. Discuss principles

More information

A novel approach to automatic model-based test case generation

A novel approach to automatic model-based test case generation Scientia Iranica D (2017) 24(6), 3132{3147 Sharif University of Technology Scientia Iranica Transactions D: Computer Science & Engineering and Electrical Engineering www.scientiairanica.com A novel approach

More information

On Constraint Problems with Incomplete or Erroneous Data

On Constraint Problems with Incomplete or Erroneous Data On Constraint Problems with Incomplete or Erroneous Data Neil Yorke-Smith and Carmen Gervet IC Parc, Imperial College, London, SW7 2AZ, U.K. nys,cg6 @icparc.ic.ac.uk Abstract. Real-world constraint problems

More information

Example: Map coloring

Example: Map coloring Today s s lecture Local Search Lecture 7: Search - 6 Heuristic Repair CSP and 3-SAT Solving CSPs using Systematic Search. Victor Lesser CMPSCI 683 Fall 2004 The relationship between problem structure and

More information

Dynamic Symmetry Breaking in Constraint Programming and Linear Programming Hybrids

Dynamic Symmetry Breaking in Constraint Programming and Linear Programming Hybrids Dynamic Symmetry Breaking in Constraint Programming and Linear Programming Hybrids Karen E. Petrie and Barbara Smith and Neil Yorke-Smith Abstract. Symmetry in Constraint Satisfaction Problems (CSPs) can

More information

Generating Implied Boolean Constraints via Singleton Consistency

Generating Implied Boolean Constraints via Singleton Consistency Generating Implied Boolean Constraints via Singleton Consistency Roman Barták Charles University in Prague, Faculty of Mathematics and Physics Malostranské nám. 2/25, 118 00 Praha 1, Czech Republic roman.bartak@mff.cuni.cz

More information

Daniel Frost and Rina Dechter. fdfrost,

Daniel Frost and Rina Dechter. fdfrost, Look-ahead value ordering for constraint satisfaction problems Daniel Frost and Rina Dechter Dept. of Information and Computer Science University of California, Irvine, CA 92717-3425 U.S.A. fdfrost, dechterg@ics.uci.edu

More information

Constraint Solving by Composition

Constraint Solving by Composition Constraint Solving by Composition Student: Zhijun Zhang Supervisor: Susan L. Epstein The Graduate Center of the City University of New York, Computer Science Department 365 Fifth Avenue, New York, NY 10016-4309,

More information

perform well on paths including satellite links. It is important to verify how the two ATM data services perform on satellite links. TCP is the most p

perform well on paths including satellite links. It is important to verify how the two ATM data services perform on satellite links. TCP is the most p Performance of TCP/IP Using ATM ABR and UBR Services over Satellite Networks 1 Shiv Kalyanaraman, Raj Jain, Rohit Goyal, Sonia Fahmy Department of Computer and Information Science The Ohio State University

More information

Department of Computer Science COMP The Programming Competency Test

Department of Computer Science COMP The Programming Competency Test The Australian National University Faculty of Engineering & Information Technology Department of Computer Science COMP1120-2003-01 The Programming Competency Test 1 Introduction The purpose of COMP1120

More information

Constraint Satisfaction. AI Slides (5e) c Lin

Constraint Satisfaction. AI Slides (5e) c Lin Constraint Satisfaction 4 AI Slides (5e) c Lin Zuoquan@PKU 2003-2018 4 1 4 Constraint Satisfaction 4.1 Constraint satisfaction problems 4.2 Backtracking search 4.3 Constraint propagation 4.4 Local search

More information

Outline. Computer Science 331. Information Hiding. What This Lecture is About. Data Structures, Abstract Data Types, and Their Implementations

Outline. Computer Science 331. Information Hiding. What This Lecture is About. Data Structures, Abstract Data Types, and Their Implementations Outline Computer Science 331 Data Structures, Abstract Data Types, and Their Implementations Mike Jacobson 1 Overview 2 ADTs as Interfaces Department of Computer Science University of Calgary Lecture #8

More information

Hybrid Algorithms for SAT. Irina Rish and Rina Dechter.

Hybrid Algorithms for SAT. Irina Rish and Rina Dechter. To Guess or to Think? Hybrid Algorithms for SAT Irina Rish and Rina Dechter Information and Computer Science University of California, Irvine fdechter,irinarg@ics.uci.edu http://www.ics.uci.edu/f~irinar,~dechterg

More information

Implementations of Dijkstra's Algorithm. Based on Multi-Level Buckets. November Abstract

Implementations of Dijkstra's Algorithm. Based on Multi-Level Buckets. November Abstract Implementations of Dijkstra's Algorithm Based on Multi-Level Buckets Andrew V. Goldberg NEC Research Institute 4 Independence Way Princeton, NJ 08540 avg@research.nj.nec.com Craig Silverstein Computer

More information

Abstract Studying network protocols and distributed applications in real networks can be dicult due to the need for complex topologies, hard to nd phy

Abstract Studying network protocols and distributed applications in real networks can be dicult due to the need for complex topologies, hard to nd phy ONE: The Ohio Network Emulator Mark Allman, Adam Caldwell, Shawn Ostermann mallman@lerc.nasa.gov, adam@eni.net ostermann@cs.ohiou.edu School of Electrical Engineering and Computer Science Ohio University

More information

Conflict based Backjumping for Constraints Optimization Problems

Conflict based Backjumping for Constraints Optimization Problems Conflict based Backjumping for Constraints Optimization Problems Roie Zivan and Amnon Meisels {zivanr,am}@cs.bgu.ac.il Department of Computer Science, Ben-Gurion University of the Negev, Beer-Sheva, 84-105,

More information

A Hybrid Recursive Multi-Way Number Partitioning Algorithm

A Hybrid Recursive Multi-Way Number Partitioning Algorithm Proceedings of the Twenty-Second International Joint Conference on Artificial Intelligence A Hybrid Recursive Multi-Way Number Partitioning Algorithm Richard E. Korf Computer Science Department University

More information

: Principles of Automated Reasoning and Decision Making Midterm

: Principles of Automated Reasoning and Decision Making Midterm 16.410-13: Principles of Automated Reasoning and Decision Making Midterm October 20 th, 2003 Name E-mail Note: Budget your time wisely. Some parts of this quiz could take you much longer than others. Move

More information

instruction fetch memory interface signal unit priority manager instruction decode stack register sets address PC2 PC3 PC4 instructions extern signals

instruction fetch memory interface signal unit priority manager instruction decode stack register sets address PC2 PC3 PC4 instructions extern signals Performance Evaluations of a Multithreaded Java Microcontroller J. Kreuzinger, M. Pfeer A. Schulz, Th. Ungerer Institute for Computer Design and Fault Tolerance University of Karlsruhe, Germany U. Brinkschulte,

More information