C. L. Forgy. The OPS83 Report. Technical Report CMU-CS-84-133, CarnegieMellon University, Department of Computer Science, 1984.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....with confidence that you have an action process verb and it needs an agent, but you don t know what that agent is, then suggest various reasons why you might not know with appropriately low confidence in them. 2The matchef is a slightly altered form of the RETE Matchef written by Forgy for OPS4 [3]. 161 Coordination of Mind and Eye The basic method of coordinating eye and mind in the present model is to make getting the next word contingent upon having completed the processing on the present one. In a production system architecture, this simply means that the match fails to turn up any ....

Forgy, C. L. OPS4 User's Manual. Department of Computer Science, Carnegie-Mellon University, 1979.

....that tell how fast their execution can be. We evaluate the rule based system in terms that are independent of the hardware platform and the execution engine. Contrary to popular expert system languages such as OPS5 where the interpretation of the language is defined by the recognize act cycle [For81], the basic interpretation cycle of EQL is defined by fixed point convergence [CBMW93] Therefore, the response time for the EQL program can be related to the time it takes to converge to a fixed point rather than to the length of the recognize act cycle. We will use the following two metrics to ....

....as finding the state in the transition diagram for which the highest number of rules have to be fired (sequentially) to reach the fixed point (see Sec.3.3) Average complexity of enabling conditions. It has been determined that 90 of execution time of rule based system is spent for matching (see [For81]) Even if the EQL language is much simpler and the enabling conditions do not require matching but rather a simple evaluation of two valued expression, this would still affect the response time. In general, the simpler are the enabling conditions, the less time would be spent to determine which ....

[Article contains additional citation context not shown here]

C. L. Forgy. OPS5 user's manual. Technical report cmu-cs-81-135, Carnegie-Mellon University, Department of Computer Science, 1981.

....(1) of defining grammars for visual languages will be postponed to Section 3.3. The development of rulebased programming language began about 25 years ago with the construction of expert systems like the medical diagnosis system MYCIN [95] EMYCIN, the development language of MYCIN, and OPS5 [41] are two early examples of textual rule based programming languages. Early examples of visual graph rewriting languages are AMBIT G [22] and PLAN2D [28] Postponing the presentation of graph rewriting (transformation) languages to Section 3.6, our history of visual rule based programming ....

C. L. Forgy. Ops5 user's manual. Technical Report CMU-CS-81-135, Department of Computer Science, Carnegie Mellon University, 1981.

....with the same priority, including instantiations of the same rule, are selected for firing by a fair nondeterministic policy. For historic reasons the current implementation dictates that fairness be defined such that a rule instantiation is fired at most once, where once has the OPS5 definition[2]. 1 This rule can be read as: for all s in r, if there does not exist a symmetric element t, then create it and add it to r. module enforcesymmetry(Relation r[ rule enforce; from r[ s; from r[ t; if( s.domainoid = t.rangeoid s.rangeoid = t.domainoid) Relation ....

C.L. Forgy. Ops5 user's manual. Technical Report CMU-CS-81-135, Department of Computer Science, Carnegie Mellon University, July 1981.

.... a number of parallel algorithms for the high speed execution of rule based systems, several having been implemented on the DADO2 parallel computer and reported in various articles [20, 18, 21] Many others have reported similar approaches and results including shared memory parallel systems [3, 5, 8, 12]. Considerable work has already been done in an effort to determine the available parallelism in rule based programs, and of the speedup to be expected from their execution on parallel machines. Studies of existing OPS5 production system programs have yielded results indicating that, in fact, a ....

Forgy C. L. OPS5 User's Manual. Technical Report CMU-CS-81-135, Department of Computer Science, Carnegie-Mellon University, July, 1981.

....of a rule may change the state of the system and cause more rules to be evaluated and fire due to this change. When the firing of rules no longer modify the current state or the modified state satisfies no rules, a Venus module is said to have reached fixed point and the module concludes execution [4]. Modules may be listed in the action of a rule and can be nested arbitrarily deeply. If a rule fires and its action lists module calls, then the rules within the nested modules must achieve fixed point before the action of the rule commits. Thus, Venus semantics and nested transaction models are ....

C.L. Forgy. OPS5 User's Manual. Technical Report CMU-CS-81-135, Department of Computer Sciences, Carnegie Mellon University, July 1981.

....matched to some extent, if not exact, a conclusion is derived considering the degree of match. Several fuzzy expert system shells have been developed to integrate the fuzzy logic into their reasoning processes. For example, FLOPS and Z II are fuzzy expert systems. FLOPS [11] an extension of OPS5 [24] by adding fuzzy concepts, selectively executes rules according to their confidence factors. Z II uses VAX Lisp, and inferences based on the compositional rule of inference [12] Recently, the implementation of Fuzzy CLIPS has been reported from NASA [25] This enables domain experts to express ....

C. L. FORGY, OPS5 user's manual, Department of Computer Science, Carnegie-Mellon University (1994).

....point of view on this integration is materialized by the N OPUS system, written in Smalltalk 80, which we will describe shortly here. This system is based on an original presentation of the OPUS system by Atkinson Laursen, in OOPSLA 87. This article described a translation of the OPS 5 system [Forgy 81] in Smalltalk 80, and contained the description of the language as well as the major implementation choices. Since the publication of this article, we have been experimenting with several versions of OPUS, eventually coming up with a new system called N OPUS. The growing success of the system in ....

....14 goals subgoals. We claim that this notion is better seen as a control problem and should be solved by proposing particular control strategies. V. Implementation Rules are compiled in N OPUS using an extended Rete network architecture. This compilation technique was pioneered by Forgy [Forgy 81] and is considered one of the best for patternmatching with many objects (it is for instance, the core of the CLIPS system [CLIPS 92] made by the NASA for real time expert systems) It was extended to take full fledged objects into account by Atkinson Laursen, in the OPUS system. The ....

Forgy, C.L. OPS-5 User Manual. Department of Computer Science, CarnegieMellon University, (1981).

....construction of a continuously operating expert system. Based on his experiences the forward chaining inference strategy is usually assumed most suitable for these systems. The state of the art of the OPS production system family, based on the RETE algorithm (Forgy 1982) is represented by OPS83 (Forgy 1984). YES L1 (Milliken 1985) is a rule based extension of PL 1. Hexscon (Wright 1986) and Escort (Sachs 1986) are more rigid languages than what we aim for, while PICON (Moore 1985) is large and specialized to continuous processes. Stimulus (Robertson 1985) has many objectives in common with us. See ....

Forgy, C.L., The OPS83 Report. Report CMU-CS-84-133, Carnegie-Mellon University, Department of Computer Science, 1984.

....in section 3. The design and implementation of DFLOPS is presented in detail in section 4. Section 5 gives an example showing how this machine works. The results evaluation is discussed in section 6. Finally, the conclusion and remarks are given. 2 Production Systems A production system[3, 7] is defined by a set of rules, or productions, which form the production memory (PM) together with a database of assertions, called working memory (WM) and an inference engine which executes the rules with a cycle consisting of three action states: 1. Match: the interpreter identifies all the ....

....satisfied by the current contents of the WM. 2. Select: the interpreter applies some conflict resolution strategies to determine one dominant rule for execution. 3. Execute: the interpreter executes the rule selected and the working memory is modified by the action part of the selected rule. OPS5[3, 7] supports forward chaining mechanism and LEX or MEA conflict resolution strategies to select the best rule to fire. Each unit of WM, consisting of a class name and several attribute value pairs, is called working memory elements(WMEs) Each WME is associated with an integer called time tag. Time ....

C. L. Forgy. OPS5 user's manual. Technical Report CMU-CS-81-135, Department of Computer Science, Carnegie-Mellon University, July 1981.

....(1) of defining grammars for visual languages will be postponed to Section 1.3. The development of rulebased programming language began about 25 years ago with the construction of expert systems like the medical diagnosis system MYCIN [46] EMYCIN, the development language of MYCIN, and OPS5 [47] are two early examples of textual rule based programming languages. Early examples of visual graph rewriting languages are AMBIT G [20] and PLAN2D [21] Postponing the presentation of graph rewriting (transformation) languages to Section 1.6, our history of visual rule based programming languages ....

C. L. Forgy. Ops5 user's manual. Technical Report CMU-CS-81-135, Department of Computer Science, Carnegie Mellon University, 1981.

.... Rule instances are collected to form the conflict set, to which selection operations are applied in order to choose a rule whose actions are to be executed (or fired) The ELEKTRA interpreter can be used as a normal forward chaining interpreter of a kind similar to the OPS family (e.g. [5]) it can be made to perform conflict resolution in the usual way. This mode of operation is a default mode; it is considered to be less important than the mode in which the interpreter executes rules at different levels of a hierarchy starting with object level rules and including a potentially ....

....wish. The temporal qualification is necessary in the last sentence because it is possible to switch between different interpreter processes in ELEKTRA, an interpreter being just a set of rules. The interpretation of the condition part of a production rule in a conventional interpreter such as OPS5 [5] can be accurately described as having a declarative semantics. In ELEKTRA, this cannot be stated because of the possibility that some elements have a procedural interpretation. In an identical fashion, the semantics of the action parts of rules is amenable to a dual interpretation. Here, ....

Forgy, C. L., The OPS5 User's Manual, Technical Report No. TR CMU-CS-81-135, Department of Computer Science, Carnegie-Mellon University, 1981.

....to attempt the construction of a continuously operating expert system. Based on his experiences the forward chaining inference strategy is usually assumed most suitable for these systems. The state of the art of the OPS family, using the RETE algorithm, Brownston 85] is represented by OPS83 [Forgy 84] It has been built so as to be tailorable and it could be used in embedded applications. See [ Arzen 86] for a use of OPS systems for process control. YES L1 [Milliken 85] a programming language descending from the YES MVS project [Griesmer 84] is the language most similar to XC. Hexscon ....

Forgy, C.L., The OPS83 report. Report CMU-CS-84-133, Carnegie-Mellon University, Department of Computer Science, 1984.

....typical applications will be highly modular with rather small rule sets. 2.2 Data Representation Production systems usually have some predefined and rather restricted way for representing working memory elements, i.e. the data used by the system. The classes of attribute value pairs in OPS5 [Forgy 81] are one such example. It is rather obvious that no single mechanism is suitable for all applications. Try, for example, to represent bit vectors with OPS5 classes. Furthermore, the internal representation of working memory elements is often complicated and direct access to the data is ....

Forgy, C.L., OPS5 user's manual. Report CMU-CS-81-135, Carnegie-Mellon University, Department of Computer Science, 1981.

....while the second formalism, logic programming, is an example of the latter. Production Systems: The production system (PS) model is perhaps the most widely used paradigm for knowledge based expert system construction. This model is supported through production system rule languages such as OPS5 [26, 27]. A PS program consists of a set of rules and a set of facts (working memory or WM) A rule consists of a left hand side (LHS) which is a conjunction of condition elements or patterns that may match elements of the WM. The right hand side (RHS) of a rule consists of actions that modify the WM. ....

....15 which corresponds to a rule instance. Thus, production rules are of the form: Patterns ) Actions where the LHS patterns are often called the condition part, and the RHS is called the action part of the rule. The operational semantics of conventional production system languages such as OPS5 [27] calls for the selection of a single instance from the set of instances formed at any given inference cycle, by applying some selection strategy (also known as the conflict resolution method) and executing the actions associated with that instance, possibly modifying the WM. This procedure is ....

[Article contains additional citation context not shown here]

C. L. Forgy. OPS5 user's manual. Technical Report CMU-CS-81-135, Department of Computer Science, Carnegie-Mellon University, July 1981. 228

....1993] simplification of many to many relationships between objects because of large branching factors, and creation of new objects versus field replication to improve access times. The rules for deriving the changes in data structures, were written in OPS5, a production system developed at CMU [Forgy 1981]. 3.2. User Interfaces Modifications to the user interface are intended to provide a more effective human computer interaction for database operations. Different users may have very different views of the data, individual interfaces need not be the same, instead they should evolve as a result of ....

Forgy, Charles L. OPS5 User's Manual. Department of Computer Science, Carnegie Mellon University.

....unavoidable serial constructs, for example, I O operations and trivial iteration of operations over working memory elements. We argue that such operations are not essentially knowledge based and are best performed using imperative constructs such as those provided in the OPS83 programming language[Forgy 1984]. One of the most problematic algorithmic constructs is that of program sequencing, that is, ensuring that a computation proceeds through a number of discrete processing phases. This is usually done by means of mode or goal working memory elements. Each rule in a given processing phase contains a ....

Forgy, C.L., The OPS83 Report, Technical Report CMU-CS-84-133, Department of Computer Science, Carnegie-Mellon University, May 1984.

....serializable and exclusive modes) implies locks gained during condition evaluation must be maintained through action execution. Thus, for VenusDB programs, a decoupled CA coupling mode is not appropriate. Further, the Venus system (as well as other rule based environments like CLIPS or OPS [5,4]) operates with individual rules as atomic elements, implying action execution immediately following condition evaluation. Thus, an immediate CA coupling mode is dictated. This leaves the EC coupling mode for chained rules. Recall that the event triggering the rule evaluation necessarily must be ....

C.L. Forgy. OPS5 User's Manual. Technical Report CMU-CS-81-135, Department of Computer Sciences, Carnegie Mellon University, July 1981.

.... learning system designers, can only avoid with certainty by having sufficient knowledge about the solution a priori It is interesting to note that all of these problems seem to disappear if we consider using a sufficiently general representation scheme (for example, OPS type production systems [For81]) but this comes at the expense of vastly increasing the size of the search space. At present, there exists no detailed analysis from the machine learning perspective of this trade off between solution set size and search space size. A better understanding of this issue would contribute a great ....

C.L. Forgy, OPS5 User's Manual, Department of Computer Science, Carnegie-Mellon University, 1981.

No context found.

C. L. Forgy. The OPS83 Report. Technical Report CMU-CS-84-133, CarnegieMellon University, Department of Computer Science, 1984.

No context found.

C. L. Forgy. OPS5 User's Manual. Technical Report CMU-CS-81-135, CarnegieMellon University, Department of Computer Science, 1981.

No context found.

C. L. Forgy. The OPS83 Report. Technical Report CMU-CS-84-133, CarnegieMellon University, Department of Computer Science, 1984.

No context found.

C. L. Forgy. OPS5 User's Manual. Technical Report CMU-CS-81-135, CarnegieMellon University, Department of Computer Science, 1981.

No context found.

C.L. Forgy: OPS5 users's manual. Report CMU-CS-81-135, Carnegie-Mellon University, Department of Computer Science, 1981.

No context found.

Forgy, C.L., The OPS83 Report. Report CMUCS -84-133, Carnegie-Mellon University, Department of Computer Science, 1984.

First 50 documents

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