A. Borning, R. Anderson, and B. Freeman-Benson. Indigo: a local propagation algorithm for inequality constraints. In Proc. 1996.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....so they can only be applied to special sets of constraints and comparators. The local propagation algorithms DeltaBlue [13] SkyBlue [12] QuickPlan [15] DETAIL [10] Houria [8] can solve only equality constraints, e.g. linear equations over reals. The exception is the Indigo algorithm [3] for solving inequalities that combines local propagation and refining method. We borrowed the main idea behind the Indigo algorithm, i.e. the propagation of the set of values, to our framework. The local propagation algorithms also use the locally predicate comparator or its variant only. Only ....

Borning, A., Anderson, R., Freeman-Benson, B., Indigo: A Local Propagation Algorithm for Inequality Constraints, in: Proceedings of the 1996.

....comparator was proposed as a method to solve the hierarchy efficiently using local comparator only. So, we can apply a local propagation algorithm for solving the hierarchy of constraints which could be easily adapted to ordered better comparator. These requirements satisfy the Indigo algorithm [BAFB96a, BAFB96b] which efficiently manipulates the acyclic set of inequality constraints. 11 The key idea in Indigo is that lower and upper bounds on variables (i.e. intervals) are propagated, and the constraints are processed from strongest to weakest, tightening the bounds on variables using ....

Alan Borning, Richard Anderson, and Bjorn Freeman-Benson. Indigo: A local propagation algorithm for inequality constraints. In Proceedings of the 1996 ACM Symposium on User Interface Software and Technology, pages 129--136, 1996.

....of some of the motion parameters (location, velocity) at certain times (Cohen 1992) In these cases, constraints are used to generate a piece of motion, which is a more limited usage than in our case. Finally, the user interface of our animation editor brings into mind graphical user interfaces (Borning et al. 1995, 1996, 1998) layout systems (Oster et al. 1998, Vander Zanden et al. 1995, Pachet et al. 1998) and systems supporting musical composition (Pachet 1999) which apply constraints and often use some specific incremental propagation method to update solutions. What makes our animation editor basically ....

Borning, A., Anderson, R., Freeman-Benson, B. (1996) Indigo: A local propagation algorithm for inequality constraints, Proc. of the ACM Symposium on User Interface Software and Technology , pp. 129-136.

....There are three main limitations of DeltaBlue: 1) it can handle only functional constraints which compute a single value for a variable (e.g. it cannot manage inequalities) 2) it cannot solve cyclic constraint graphs; and 3) all methods must have exactly one output variable. The Indigo solver [11] relaxes the first restriction by propagating bounds on value assignments instead of specific values Indigo binds variables to intervals. This generalization requires the solver to fire multiple interval tightening methods instead of just a single method performing a value assignment. Thus, if ....

....real numbers can have an associated error function. The error in satisfying a constraint is 0 if and only if the constraint is satisfied, and becomes larger the less nearly satisfied the constraint is. For inequality constraints it is important to use a metric rather than a predicate comparator [11]. Thus, plausible comparators for use with linear equality and inequality constraints are 43 locally error better, weighted sum better,andleast squares better. For a given collection of constraints, Cassowary finds a locally error better or a weighted sum better solution. The related QOCA ....

[Article contains additional citation context not shown here]

Alan Borning, Richard Anderson, and Bjorn Freeman-Benson. Indigo: A local propagation algorithm for inequality constraints. In Proceedings of the 1996 ACM Symposium on User Interface Software and Technology, pages 129--136, Seattle, November 1996. 155

....of some of the motion parameters (location, velocity) at certain times (Cohen 1992) In these cases, constraints are used to generate a piece of motion, which is a more limited usage than in our case. Finally, the user interface of our animation editor brings into mind graphical user interfaces (Borning et al. 1995, 1996, 1998) layout systems (Oster et al. 1998, Vander Zanden et al. 1995, Pachet et al. 1998) and systems supporting musical composition (Pachet 1999) which apply constraints and often use some specific incremental propagation method to update solutions. What makes our animation editor basically ....

Borning, A., Anderson, R., Freeman-Benson, B. (1996) Indigo: A local propagation algorithm for inequality constraints, Proc. of the ACM Symposium on User Interface Software and Technology , pp. 129-136.

....was proposed as a method to solve the hierarchy ef Thetaciently using local comparator only. So, we can apply a local propagation algorithm for solving the hierarchy of constraints which could be easily adapted to ordered better comparator. These requirements satisfy the Indigo algorithm [BAFB96a, BAFB96b] which ef Thetaciently manipulates the acyclic set of inequality constraints. The key idea in Indigo is that lower and upper bounds on variables (i.e. intervals) are propagated, and the constraints are processed from strongest to weakest, tightening the bounds on variables using ....

Alan Borning, Richard Anderson, and Bjorn Freeman-Benson. Indigo: A local propagation algorithm for inequality constraints. In Proceedings of the 1996 ACM Symposium on User Interface Software and Technology, pages 129#136, 1996.

....and inequalities, neither of which is supported by traditional local propagation algorithms. Examples include QOCA [58] which solves simultaneous linear equation and inequality constraints while optimizing a quadratic expression, Bramble [46] and Juno 2 [65] which use numerical solvers, Indigo [9], an interval propagation algorithm for inequality constraints, and DETAIL [67] and Ultraviolet [10] both of which are hybrid algorithms supporting both local propagation and cycle solvers. 3.4 Constraint Programming in Operations Research Operations Research is a vast field represented by ....

A. Borning, R. Anderson, and B. Freeman-Benson. Indigo: A local propagation algorithm for inequality constraints. In Proceedings of the ACM SIGGRAPH Symposium on User Interface Software and Technology, Seattle, November 1996.

....i = v i (2.1) u g v j 8i:u i v i 9i 0 : u i 0 v i 0 : 2.2) Locally better using the predicate error function is called locally predicatebetter , and the one with the metric function is called locally metric better . 3 3 Locally metric better is also known as locally error better [5]. CHAPTER 2. CONSTRAINT HIERARCHIES 15 Locally better considers each constraint individually. It is often unable to compare variable assignments because of the situation that one assignment produces an error smaller than the other for some constraint but larger for another constraint. For ....

....special treatment of a cycle, which is considered to degenerate its performance. In addition to the demand for addressing such problems, we are now encountering a new trend that constraint hierarchy algorithms are differentiated into two levels: meta algorithms [10] and specialized algorithms [5]. A meta algorithm maintains the whole constraint hierarchy. Intuitively, it divides the hierarchy into a set of sub hierarchies based on its graph topology, making appropriate specialized algorithms actually solve the subhierarchies. Since developing an efficient, general algorithm is quite ....

[Article contains additional citation context not shown here]

Borning, A., R. Anderson, and B. Freeman-Benson, "Indigo: A Local Propagation Algorithm for Inequality Constraints," in Proceedings of the ACM Symposium on User Interface Software and Technology (UIST), Nov. 1996, pp. 129--136.

....with different input parameters. We can view this as a kind of partial evaluation of the constraint solving algorithm. This has long been done for local propagation solvers (e.g. 1] and more recently for simultaneous linear equations [4] and for acyclic sets of inequality constraints [2]. However, there have not been any systems that can compile plans for systems of constraints including a cyclic set of simultaneous equalities and inequalities. That lack is addressed here. The original motivation for this work was constraint solving for user interface toolkits and other kinds of ....

Alan Borning, Richard Anderson, and Bjorn Freeman-Benson. Indigo: A local propagation algorithm for inequality constraints. In Proceedings of the

....with different input parameters. We can view this as a kind of partial evaluation of the constraint solving algorithm. This has long been done for local propagation solvers (e.g. 1] and more recently for simultaneous linear equations [3] and for acyclic sets of inequality constraints [2]. However, there have not been any systems that can compile plans for systems of constraints including a cyclic set of simultaneous equalities and inequalities. That lack is addressed here. The original motivation for this work was constraint solving for user interface toolkits and other kinds of ....

Alan Borning, Richard Anderson, and Bjorn Freeman-Benson. Indigo: A local propagation algorithm for inequality constraints. In Proceedings of the 1996 ACM Symposium on User Interface Software and Technology, pages 129-136, Seattle, November 1996.

....real numbers can have an associated error function. The error in satisfying a constraint is 0 if and only if the constraint is satisfied, and becomes larger the less nearly satisfied the constraint is. For inequality constraints it is important to use a metric rather than a predicate comparator [Borning et al. 1996]. Thus, plausible comparators for use with linear equality and inequality constraints are locally error better, weighted sum better, and least squares better. For a given collection of constraints, Cassowary finds a locally error better or a weighted sum better solution. The related QOCA ....

.... Sutherland s pioneering Sketchpad system [Sutherland 1963] Most of the current systems use one way constraints (e.g. Hudson and Smith 1996; Myers 1996] or local propagation algorithms for acyclic collections of multi way constraints (e.g. Sannella et al. 1993; Vander Zanden 1996] Indigo [Borning et al. 1996] handles acyclic collections of inequality constraints, but not cycles (simultaneous equations and inequalities) User interface systems that handle simultaneous linear equations include DETAIL [Hosobe et al. 1996] and Ultraviolet [Borning and Freeman Benson 1995] A number of researchers ....

[Article contains additional citation context not shown here]

Borning, A., Anderson, R., and Freeman-Benson, B. 1996. Indigo: A local propagation algorithm for inequality constraints. In Proceedings of the 1996 ACM Symposium on User Interface Software and Technology (Seattle, Nov. 1996), pp. 129--136.

....propagation when possible, and grouping constraint cycles into cells, which are then solved by an appropriate subsolver. The most recent version of DETAIL [19] also includes an experimental solver for inequality constraints as well as functional constraints. Ultraviolet uses the Indigo algorithm [1] for acyclic collections of inequality constraints. Indigo is an interval propagation algorithm; there has been considerable work on interval constraints in other areas of computer science, particularly artificial 4 intelligence and constraint logic programming. Davis [9] discusses the ....

....queue on these same variables (thus causing tightenings to ripple out through the constraint graph) This continues until the queue is empty. We keep track of any variables whose bounds have been tightened in the changed variables set. A complete description of the algorithm is given in reference [1]; proofs of correctness theorems are given in [2] 5.3. Purple Purple solves collections of constraints that can be represented as linear equations. Unlike Blue and Indigo, Purple is not troubled by cycles (i.e. simultaneous equations) The algorithm is adapted from that used in CLP(R) 26] and ....

Alan Borning, Richard Anderson, and Bjorn Freeman-Benson. Indigo: A local propagation algorithm for inequality constraints. In Proceedings of the 1996 ACM Symposium on User Interface Software and Technology, pages 129--136, Seattle, November 1996.

....it is (a metric comparator. Constraints whose domain is a metric space, for example the reals, can have an associated error function. The error in satisfying a constraint cn is 0 iff the constraint is satisfied, and becomes larger the less nearly satisfied is the constraint. As described in [2], for inequality constraints it is important to use a metric rather than a predicate comparator. Thus, plausible comparators for use with linear equality and inequality constraints are locally error better, weighted sum better, and least squares better. For a given collection of constraints, ....

.... constraints in user interfaces and interactive systems, beginning with Ivan Sutherland s pioneering Sketchpad system [19] Most of the current systems use one way constraints (e.g. 12, 16] or local propagation algorithms for acyclic collections of multi way constraints (e.g. 18, 20] Indigo [2] handles acyclic collections of inequality constraints, but not cycles (simultaneous equations and inequalities) UI systems that handle simultaneous linear equations include DETAIL [11] and Ultraviolet [3] A number of researchers (including the first author) have experimented with a ....

Alan Borning, Richard Anderson, and Bjorn Freeman-Benson. Indigo: A local propagation algorithm for inequality constraints. In Proceedings of the 1996 ACM Symposium on User Interface Software and Technology, pages 129--136, Seattle, November 1996.

....(Constraints whose domain is a metric space, for example the reals, can have an associated error function. The error in satisfying a constraint cn is 0 if and only if the constraint is satisfied, and becomes larger the less nearly satisfied is the constraint. As recognized for the Indigo solver [2], for inequality constraints it is important to use a metric rather than a predicate comparator. Thus, plausible comparators for use with linear equality and inequality constraints are locally error better, weighted sum better, and least squares better. For a given collection of constraints, ....

.... constraints in user interfaces and interactive systems, beginning with Ivan Sutherland s pioneering Sketchpad system [18] Most of the current systems use oneway constraints (e.g. 11, 15] or local propagation algorithms for acyclic collections of multiway constraints (e.g. 17, 19] Indigo [2] handles acyclic collections of inequality constraints, but not cycles (simultaneous equations and inequalities) UI systems that handle simultaneous linear equations include DETAIL [10] and Ultraviolet [3] A number of researchers (including the second author) have experimented with a ....

[Article contains additional citation context not shown here]

Alan Borning, Richard Anderson, and Bjorn Freeman-Benson. Indigo: A local propagation algorithm for inequality constraints. In Proceedings of the 1996 ACM Symposium on User Interface Software and Technology, pages 129--136, Seattle, November 1996.

....be satisfied. The other strengths all label non required constraints. A constraint of a given strength completely dominates any constraint with a weaker strength. In the theory, a comparator is used to compare different possible solutions to the constraints and select among them. As described in [2], it is important to use a metric rather than a predicate comparator for inequality constraints. Thus, plausible comparators for use with linear equality and inequality constraints are locally error better, weighted sum better, and least squares better. The least squares better comparator ....

.... constraints in user interfaces and interactive systems, beginning with Ivan Sutherland s pioneering Sketchpad system [20] Most of the current systems use one way constraints (e.g. 13, 17] or local propagation algorithms for acyclic collections of multi way constraints (e.g. 19, 21] Indigo [2] handles acyclic collections of inequality constraints, but not cycles (simultaneous equations and inequalities) UI systems that handle simultaneous linear equations include DETAIL [12] and Ultraviolet [3] A number of researchers (including the first author) have experimented with a ....

A. Borning, R. Anderson, and B. Freeman-Benson. Indigo: A local propagation algorithm for inequality constraints. In UIST'96, pages 129--136, Seattle, Nov 1996.

....with different input parameters. We can view this as a kind of partial evaluation of the constraint solving algorithm. This has long been done for local propagation solvers (e.g. 1] and more recently for simultaneous linear equations [3] and for acyclic sets of inequality constraints [2]. However, there have not been any systems that can compile plans for systems of constraints including both simultaneous equalities and inequalities. That lack is addressed by the research reported here. In brief, our algorithm works as follows. The original set of constraints is converted into a ....

....pioneering Sketchpad system [14] 1 Bjorn Freeman Benson, Object Technology International, Personal Communication. Most of the current systems use one way constraints (e.g. 9, 10] or local propagation algorithms for acyclic collections of multi way constraints (e.g. 12, 15] Indigo [2] handles acyclic collections of inequality constraints, but not cycles. UI systems that handle simultaneous (cyclic) linear equations include DETAIL [8] and Ultraviolet [3] UI systems that handle simultaneous linear inequalities as well with reasonable efficiency are QOCA [7] and Cassowary [6] ....

A. Borning, R. Anderson, and B. Freeman-Benson. Indigo: A local propagation algorithm for inequality constraints. In Procs. ACM Symp. on User Interface Software and Technology, 129--136, Seattle, 1996.

No context found.

A. Borning, R. Anderson, and B. Freeman-Benson. Indigo: a local propagation algorithm for inequality constraints. In Proc. 1996.

No context found.

A. Borning, R. Anderson, and B. Freeman-Benson. Indigo: A local propagation algorithm for inequality constraints. In Proceedings of the 1996.

No context found.

Borning, A., Anderson, R., and Freeman-Benson, B. "Indigo: A local propagation algorithm for inequality constraints", In Proceedings of the ACM SIGGRAPH Symposium on user interface software and technology , Seattle, Nov. 1996.

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