Sussman, G. J., and G. L. Steele Jr., 1980. CONSTRAINTS---A language for expressing almosthierarchical descriptions, Artificial Intelligence 14, 1-39.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....than rule out impossible construals, our approach attempts to rule in the most probable construal. To arrive at this construal, TNT reasons about the topology of the input system via a rich vocabulary for representing locality. This use of locality is similar in spirit to Sussman s slice concept [50], which used multiple overlapping local viewpoints to analyze electrical circuits. Davis [8] has also noted that multiple representations of locality can provide inferential leverage in diagnostic applications. The evidential reasoning algorithm employs Pearl s [37] method for Bayesian ....

....parsing that enables the reasoner to proceed directly from structure to a functional construal of the input. We consider this to be one of the main contributions of this work. TNT s multiple representations of locality are similar in spirit to the view of locality in Sussman s slice concept [50]. Whereas the TNT reasoning process is purely topological, disregarding the geometry of the input system, Davis [8] argues that a useful definition of locality arises from an explicit representation of physical organization. For diagnostic tasks, he notes that aspects of a system that are ....

G.J. Sussman, G.L. Steele, CONSTRAINTS---A language for expressing almost-hierarchical descriptions, Artificial Intelligence 14 (1) (1980) 1--39.

....We believe that the ability to draw important inferences about a design problem early in the process is important in concurrent design. A large body of research exists on solving constraint propagation problems including that of Sutherland [39] Mackworth [20] Borning [3] Sussman and Steele [38], Gosling [14] Popplestone [27] Steward [36] Sadeh [31] and Serrano [34] These techniques provide a core of solution methods directly applicable to algebraic constraints in real variables. Based on these methods, we are developing propagation techniques applicable to constraints among ....

Sussman, G. J. and Steele, G. L., "CONSTRAINTS-- A Language for Expressing Almost Hierarchical Constraints," Artificial Intelligence, No. 14, 1980, pp. 1-39.

....In the limit, all values are precisely specified (i.e. all variables take on real values) In that case, our results match exactly with those specified by any traditional circuit analyzer. 6. Related Work There have been many successful effort to perform model based reasoning about circuits [18, 19, 4, 3, 11, 13, 6, 8, 16]. The most recent work and more related to this one are the works of Flores and Farley, and Mauss. The work of Flores and Farley can only cope with series parallel reducible circuits, missing an important number of circuits and being of no much use for many practical applications. For instance, to ....

G. J. Sussman and G. L. Steele. CONSTRAINTS: a language for expressing almost-hierarchical descriptions. Artificial Intelligence, 14:1--39, 1980.

....reduction, its children are the elements of the star, and its ancestors, S nodes, represent the Kn(n) that replaces the star. 9 Figure 5: Final Reduction Graph for Circuit of Figure 4 5 Related Work There have been many successful effort to perform model based reasoning about circuits [SS77, SS80, de 84, Dav84, Gen84, Ham91, FF96, Flo97, Mau98] The most recent work and more related to this one are the works of Flores and Farley, and Mauss. The work of Flores and Farley can only cope with series parallel reducible circuits, missing an important number of circuits and being of no much use ....

G. J. Sussman and G. L. Steele. CONSTRAINTS: a language for expressing almost-hierarchical descriptions. Artificial Intelligence, 14:1--39, 1980.

....needs. All this, of course, must be transparent to the final user of the propagation engine, whio sees the constraints in a uniform representation. INTRODUCTION Constraint propagation is an important inference engine for different AI applications that reason about quantities. Many systems [4, 6, 9, 14, 12] represent the knowledge they reason about in terms of mathematical relations on qualitative or quantitative values of quantities. For example, in circuit analysis (the application domain we are studying [5] the circuit can be modeled in terms of the Kirchoff and ohm s laws that govern its ....

....25] but we took the intersection between that and the old value. Any value can be seen as another constraint, therefore the computed value must be consistent with the old one. This guarantees that the intervals always shrink and that the propagation procedure will converge for linear functions [12]. Given the application domain we are dealing with is linear circuits in sinusoidal steady state (see section Applications) the values the variables take on belong to the complex domain, which we chose to represent in polar form. Since we want our reasoner to work with values at different ....

[Article contains additional citation context not shown here]

G. J. Sussman and G. L. Steele. CONSTRAINTS: a language for expressing almosthierarchical descriptions. Artificial Intelligence, 14:1--39, 1980.

....propagation. Also, the confluence model was extended to cope with vectors expressed in polar form. 70 CHAPTER III HYBRID REPRESENTATION CONSTRAINT PROPAGATION Constraint propagation is an important inference engine for a variety of AI applications that reason about quantities. Many systems [12, 17, 27, 50, 43] represent the knowledge they reason about in terms of mathematical relations on qualitative or quantitative values of quantities. As mentioned in the previous section, we model electrical circuits as a set of constraints and perform constraint propagation to infer as much (qualitative and or ....

....computed value must be consistent with the old one. This guarantees that the intervals always shrink 1 This feature is not present in imperative languages, where the order of computation is fixed by the program designer. 82 and that the propagation procedure will converge for linear functions [43]. Our implementation of value propagation maintains a list of variables, associating with each variable a value and a list of expressions in which it is found. For the previous example, A would contain the value A = 0; 20] and the two expressions B = A Gamma C, and C = A Gamma B. When we ....

[Article contains additional citation context not shown here]

G. J. Sussman and G. L. Steele. CONSTRAINTS: a language for expressing almost-hierarchical descriptions. Artificial Intelligence, 14:1--39, 1980.

....highly specialized nature of the constraint based programming systems available. For example, the first constraint based system, Sketchpad [29] was oriented towards graphics while Thinglab [6] was developed for simulation applications. An early language which o#ered more generality was Constraints [28] but several important notions, including inequality, were not available in the language. Magritte [14] was restricted to algebraic relationships and did not support the use of arbitrary application specific predicates, functions and domains. More recently, there has been a surge of interest in ....

Sussman, G. and Steele, G. (1980) Constraints: A Language for Expressing Almost-Hierarchical Descriptions. Artificial Intelligence, 14, 1-39.

....in Arti cial Intelligence. 1 solving algorithms for various classes of problems. Constraint programming has a long tradition in arti cial intelligence. It can be traced back to the use of constraints in Sutherland s SKETCHPAD [67] the CONSTRAINT programming language of Sussman and Steele [66] and the work of Borning on ThingLab [2] among others. Mackworth also advocated, as early as 1977, the use of consistency techniques (a paradigm emerging from arti cial intelligence to solve combinatorial search problems) in declarative languages as an alternative to chronological backtracking ....

....introduced in [54] in the context of concurrent logic programming, while the cardinality combinator was proposed explicitly for CLP languages in [74] 2. 4 The Implication Combinator Motivation Local propagation is one of the key ideas behind constraint programming languages such as CONSTRAINTS [66] and ThingLab [2] Local propagation (or value propagation) amounts to deducing values for some variables from those of other variables. For instance, an and gate in a digital circuit may be de ned by rules of the form If one input is 0 then the output is 0, If the output is 1 then the ....

G.J. Sussman and G.L. Steele. CONSTRAINTS A Language for Expressing AlmostHierarchical Descriptions. AI Journal, 14(1), 1980.

....described by a set of objects, together with some relationships among them. The recognition of this fact has lead to much work in the fields of AI and of Logic Programming. In the last twenty years a number of AI researchers have explored the use of constraints to solve di#cult problems [8, 20, 21, 22]. Most of the proposed systems were based on the technique of constraint propagation over a declarative structure called constraint network [7] A constraint network consists of a number of nodes connected by constraints. A node represents an individual parameter of the problem at hand, while a ....

G. J. Sussman and G. L. Steele. Constraints: a Language for Expressing Almost Hierarchical Descriptions. Artificial Intelligence, 14(1):1--39, 1980.

....domain of real numbers is discussed 1 . In the following, a brief description of the languages CHIP and CLP(R) is given. 1.3. 1 The Constraint Logic Programming Language CHIP Constraint programming in procedural languages has been studied in the area of Artificial Intelligence for many years [SJ80, Dav87]. Constraint satisfaction problems (CSP) where variables take values over finite domains have also attracted much attention in the AI community [Mac77, HE80, Dec90, Nad87] Research in this area was stimulated by the well known thrashing behaviour of naive backtracking. An excellent discussion of ....

G.J Sussman and G.L Steele Jr. CONSTRAINTS - a Language for Expressing Almost-Hierarchical Descriptions. Artificial Intelligence, 14:1--39, 1980.

....or intersection of geometric elements) The thesis of Yu [87] studies the precision necessary for the classification problems in polyhedral modeling, but is not optimistic about exact algorithms. There is a large literature about constraint solving geometric edit19 ing modeling systems ([77,13] are two early papers) Many of them are heuristic and partial in their solution capabilities. Recent attempts at such systems aim at more complete constraint solving abilities, by exploiting the appropriate algebraic tools such as Grobner bases, Wu Ritt theory or elimination theory. For instance, ....

G. Sussman and G. Steele. Constraints -- a language for expressing almosthierarchical descriptions. Artificial Intelligence, 14:1--39, 1980.

....can be skipped. Several systems and formalisms have influenced the development of Except II. One important influence came from work at MIT in the late seventies. After MicroPlanner, a lot of subsequent systems and techniques have there been developed. This includes Steele s system Constraints [Sussman and Steele Jr. 1980], Doyle s Truth Maintenance System (TMS) Doyle, 1979] and explicit control of reasoning [de Kleer et al. 1977] Johan de Kleer later developed the ATMS (assumption based truth maintenance system) de Kleer, 1986a] At the beginning of the eighties, the need of a theoretical basis for a lot of ....

....v2) This rule computes the value of the output of an adder a1. If it is applied to (premise (has value (input1 a1) 3) premise (has value (input2 a1) 2) 25 we obtain the conclusion (has value (output a1) 5) In section 4. 2 we take such rules to formulate more constraints in the style of [Sussman and Steele Jr. 1980]. Additionally, we use attached Lisp expressions as conditions of rules: condition ) eval lisp term) As an example consider a first attempt to formulate a divisor rule: has value (input1 d1) v1) has value (input2 d1) v2) has value (output d1) eval ( v1 v2) We get in trouble if ....

[Article contains additional citation context not shown here]

G.J. Sussman and G.L. Steele Jr. CONSTRAINTS - a language for expressing almost-hierarchical descriptions. Artificial Intelligence, 14:1--39, 1980. 102

....works is that constraint solving was performed numerically using a Newton Raphson solver. Another collection of early works arose from MIT, motivated by applications in electrical circuit analysis and synthesis, gave rise to languages for general problem solving. In the CONSTRAINTS language [240], variables and constraints are static, and constraint solving was limited to using local propagation. An extension of this work [241] provided a more sophisticated environment for constraint programming, including explanation facilities. Some other related systems, EL ARS [238] and SYN [142] ....

G. Steele & G.J. Sussman, CONSTRAINTS - A Language for Expressing Almost Hierarchical Descriptions, Artificial Intelligence 14(1), 1--39, 1980.

....other aspects of design. The SID tends to cluster into subproblems. A relation among items both in and outside a cluster represents, from the point of view of that cluster, an externally imposed constraint; such constraints make the diagram an almost hierarchical constraint network in the sense of Sussman and Steele (1980). Relations and properties are generally inherited along directed paths in the task DAG. The most obvious items not inherited are aggregate properties. This inheritance takes two forms. Relations and properties are inherited in a weak sense when several descendants taken together fulfill the role ....

Sussman, G.J. and Steele, G.L., "CONSTRAINTS - A Language for Expressing Almost-Hierarchical Descriptions," Artificial Intelligence 14 (1980), pp. 1-39.

....static constraints that were solved by local propagation and relaxation techniques. Another language was THINGLAB [7] that was a somewhat object oriented graphical system, local propagation and relaxation were used to handle the static constraints. In the area of electric circuiting CONSTRAINTS [36] was a language where variables and constraints were static, i.e. could not be dynamically generated, and the constraint solving was limited to using local propagation, it was used in applications for electrical circuit analysis and synthesis. EL ARS 2 [35] and SYN [11] was related systems that ....

G. Steele and G. J. Sussman. CONSTRAINTS - a language for expressing almost hierarchical descriptions. Artificial Intelligence, 14(1):1--39, 1980.

....satisfaction problems (CSP) The study of CSP has a long history, and we mention here just a few important references. The concept of arc consistency was introduced in [Mac77, Fre78] 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] finally [Van89] extensively motivates and describes in detail the integration of finite domain propagation methods into logic programming. 3.2 Propagation in Constraint Logic Programming The idea behind local ....

G.J. Sussman and G.L. Steele. CONSTRAINTS: A language for expressing almost-hierarchical descriptions. Artificial Intelligence, 14(1):1--39, January 1980.

....constraints are used to eliminate parts that ordinarily might be chosen to implement some function, without regard to whether that part violates a constraint or set of constraints. In part selection, constraints can be used for two purposes: evaluation and propagation (Davis 1987; Freuder 1978; Sussman and Steele 1980). In evaluation, constraints determine if the current assignment of values to design attributes is physically allowed, or feasible. In propagation, constraints infer the values of unassigned design attributes, given the values of assigned design attributes, potentially restricting future ....

....scope of our problem definition. Constraints are used in a variety of ways to propagate the effects of an assignment to a design attribute to evaluate specific decisions (Bowen, O Grady and Smith 1990; Graf, Van Hentenryck, Pradelles et al. 1989; Havens and Rehfuss 1989; Murtagh and Shimura 1990; Sussman and Steele 1980). In part selection problems that use the part as the fundamental entity to describe a design, parts are selected to implement functions, and, then, the constraints are used to propagate the effects of that decision on other design attributes. If constraint violations result, a previous part ....

Sussman, G. J. and G. L. Steele: 1980, CONSTRAINTS - A Language for Expressing Almost-Hierarchical Descriptions, Artificial Intelligence 14: 1-39.

....reduction for the satisfaction of constraints on hierarchical objects; it was designed for layout of graphics in documents, and as such ran only in batch mode. The GROW system [Barth86] provided constraints as an intrinsic part of an object oriented, graphical user interface toolkit. CONSTRAINTS [Sussman80] provided for the specification of hierarchical constraint equations for electrical circuits using local propagation. CONSTRAINT [VanderZanden88] is a language with which graphical objects and subobjects may be displayed and modified, while satisfying a set of constraint equations, ordering these ....

Sussman, G. J. and Steele, G. L., Jr. CONSTRAINTS--A Language for Expressing Almost-Hierarchical Descriptions. Artificial Intelligence, 14, 1-39, 1980.

....called constraints. An assignment of one value to each variable such that every constraint is satisfied is called a solution, the search for a solution is called a constraint satisfaction problem (CSP) 10] Sussman and Steele were the first who used constraint nets to describe electric circuits [14]. To test such a circuit, an observation must be made. e.g. the input voltage U i is given and the output voltage U o is measured. Then the circuit is to be considered correct if, and only if, there exists the currents I i , I g , and I o so that the equations are simultaneously satisfied. In ....

G.J. Sussman and G.L. Steele. CONSTRAINTS --- a language for expressing almost hierarchical descriptions. Artificial Intelligence, 14:1--39, 1980.

....each variable such that each constraint (i.e. each condition) is satisfied. Constraint nets can be used as system descriptions: variables represent visible signals (i.e. sensor and control signals) and unknown internal states 1 , constraints represent connections between signals and states. Sussman and Steele (1980) have used such a system description for electric circuits without time dependent behavior. In order to describe real time systems, constraint nets must be temporally extended. In this case, values change with time, and constraints must express temporal connections between variables. In the ....

Sussman, G. and G.L. Steele (1980). CONSTRAINTS --- a language for expressing almost hierarchical descriptions. Artificial Intelligence 14, 1--39.

No context found.

Sussman, G. J., and G. L. Steele Jr., 1980. CONSTRAINTS---A language for expressing almosthierarchical descriptions, Artificial Intelligence 14, 1-39.

No context found.

Sussman, G. J., and G. L. Steele Jr., 1980. CONSTRAINTS---A language for expressing almost-hierarchical descriptions, Artificial Intelligence 14, 139. Thaler, R. H., and Shefrin, H. M., 1981. An economic theory of self-control, J. Political Economy, Vol. 89, No. 2, 392-406.

No context found.

G. J. Sussman and G. L. Steele. CONSTRAINTS: a language for expressing almost-hierarchical descriptions. Artificial Intelligence, 14:1--39, 1980.

No context found.

Sussman, Gerald Jay and Guy Lewis Steele, Jr, 1980. Constraints: A Language for Expressing Almost-Hierarchical Descriptions, Artificial Intelligence Journal, vol. 14, 1980, 1-39.

No context found.

G. J. Sussman and G. L. Steele. CONSTRAINTS--A Language for Expressing Almost-hierarchical Descriptions. Journal of Artificial Intelligence, 14 (1), 1980.

First 50 documents  Next 50

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