Showing papers on "Heuristic published in 1970"

On generality and problem solving: a case study using the DENDRAL program

Abstract: Heuristic DENDRAL is a computer program written to solve problems of inductive inference in organic chemistry This paper will use the design of Heuristic DENDRAL and its performance on different problems for a discussion of the following topics: 1 the design for generality; 2 the performance problems attendant upon too much generality; 3 the coupling of expertise to the general problem solving processes; 4 the symbiotic relationship between generality and expertness, and the implications of this symbiosis for the study and design of problem solving systems We conclude the paper with a view of the design for a general problem solver that is a variant of the "big switch" theory of generality

Note---On Madigan's Approach to the Deterministic Multi-Product Production and Inventory Problem

Abstract: The deterministic multi-product single-machine production and inventory problem is one of the simplest models in which production and inventory factors interact. There is at present no algorithm available which solves the problem optimally, and several different types of approaches have been presented in the literature. Madigan's heuristic method [Madigan, J. G. 1968. Scheduling a multi-product single machine system for an infinite planning period. Management Sci.1411, July 713--719.] avoided some of the drawbacks characteristic of other methods, but there are some errors and limitations in his original presentation, to which this note is addressed.

A heuristic approach to task dispatching

Abstract: This paper describes an experimental algorithm for allocating use of a central processing unit to perform separate data processing tasks in a multitasking system. The algorithm, which may control only a subset of the tasks being performed by the system, appears to improve run time for some work loads. Tasks with a recent history of using input/output facilities are given preference. This heuristic treatment of tasks is carried over to the algorithm itself, which modifies its own characteristics based on its overall effectiveness in handling the tasks under its control.

Production and Workforce Scheduling Extensions

Abstract: Many papers have approached the problem of determining optimum production and workforce scheduling plans. The most common approaches have been analytical, iterative, and heuristic. This paper extends work done on the analytical approach to include consideration of multiproduct facilities and products with long production cycle times.

Heuristic Search Programs

Abstract: When a certain problem cannot be solved directly, it is often possible to write a program to search for a solution If the number of possibilities to be searched in order to find a solution is sufficiently small, the problem is trivial since the program can consider all possibilities For an intellectually difficult problem, the number of possibilities to be searched is so large (sometimes infinite) that for all practical purposes there is no exhaustive procedure For example, in most kinds of theorem-proving including predicate calculus, the number of possibilities to be searched is potentially infinite In order to guarantee a perfect first move in checkers, the program would have to search through about 1040 possible games In chess the number is about 10120Even tremendous improvements in hardware would hardly dent tasks requiring the search of so many possibilities It is much better to consider alternative ways of defining the search and modifying the search, that is, to consider heuristic search techniques In general, a heuristic is a rule of thumb, strategy, method or trick used to improve the efficiency of a system which tries to discover the solutions of complex problems Some heuristics are specific, that is, limited to one problem-solving domain, such as proving theorems in geometry Other heuristics are general, that is, applicable to several domains For example, the heuristic of “working backwards” is useful in many theorem-proving and problem-solving domains

Analysis of a Decision Making Investigation

Abstract: Those management and social scientists dealing with organizational behavior are severely limited by the need for more operational theory and by the paucity of experimental data on important events and phenomena. The many practical and conceptual reasons why experiments in the classical sense cannot be conducted lead us to consider a less rigorous form of inquiry called an investigation. A possible heuristic procedure for analyzing data from investigations is proposed. The heuristic has the property that rather than selecting favorable events to support a theory, it codes and incorporates observed events systematically into a testing of theory. This procedure is applied to data from the Churchman-Ratoosh laboratory investigations of implementation processes. Our results are consistent with the assumption of a constant source diffusion of innovation process. Data for the diffusion process and attendant structural change are presented in terms of models. The application of the heuristic also suggests questio...

On applications of graph/network theory to problems in communication systems

Abstract: Graph/network theory results are applicable to problems in communications. As a representative example, the node coloring problem in graph theory is applicable to the channel assignment problem in cellular mobile communication systems. The node coloring problem is NP-complete, meaning that optimally solving it is very difficult. Therefore, we use heuristic algorithms for the channel assignment problem. In this case, the graph theory results show the legitimacy of using heuristic techniques. On the other hand, we can directly apply graph theory to communication problems. For example, on contents delivery services in the Internet, we place mirror servers that provide the same contents on the network. Location problems on flow networks are applicable to mirror server allocation problems. In a simple case, we can efficiently solve the problem. In this paper, we concentrate on multi-hop wireless networks and consider the relationship between their problems and the results of graph/network theory.

A study of heuristic learning methods for optimization tasks requiring a sequence of decisions

Abstract: Learning is a broad term covering many different phenomena. It is convenient to segment learning into three different problems in induction: the collection and use of stochastic information on past performance in order to improve performance, the determination of which variables are relevant to the decisions being made, and the derivation of performance rules in the predicate calculus from the collected data. This study concentrates on the first problem.

Heuristic interactive synthesis of cellular arrays : based on Polish notation

Abstract: The object of the research' contained in this thesis is to develop a workable design tool for embedding arbitrary Boolean switching functions in a class of cellular arrays. Such a system is produced in an interactive mode, using a heuristic approach, and based upon the notational system called Polish notation. A correspondence is developed between arbitrary Polish notation representations of Boolean functions and a class of cellular arrays. Heuristic manipulative porcedures for Polish strings are developed. These procedures are applied to reduce the Polish string to a form corresponding to a near minimal number of cells in the array realizing the function. Options available to the designer are presented, with the purpose of further reducing the array, using the gestalt forming abilities of the human operator. The class of arrays is shown to be inadequate for all switching problem formulations and a new class of arrays, the buss arrays, is introduced to overcome this deficiency. An algorithm is developed to embed Polish string representations of Boolean functions in buss arrays. A heuristic interactive approach to applying and modifying this algorithm is presented. Examples are produced, including a cellular multiplier.

B70-4 Theory of Scheduling

Abstract: Scheduling and resource allocation have come to the forefront of the problems that confront the designers of computer systems. This is particularly true in time-sharing, multiprogramming, and real-time systems, but it is also the case for somewhat less sophisticated systems due to problems created by disk accessing. Theory of Scheduling is a very readable treatment of some of the "classical" solutions to scheduling problems, as well as some heuristic ad hoc approaches that are suitable in particular computer environments. The operating systems designer should not consider this book as the panacea for all of his ills, however, because the methodology that is given for solving his scheduling problems may be unsatisfactory in the context in which he chooses to use it. For example, in a real-time system the scheduling algorithm must satisfy constraints on maximum computation time which are not necessarily satisfied by the algorithms described in the text.

A Barge Sequencing Heuristic

Abstract: This paper describes a heuristic method for sequencing barge unloadings at terminals that have regulated unloading rates or flows with the objective of minimizing the variation of a common constituent in the final combined output. The work was undertaken in response to the desire of a steel manufacturer to minimize the maximum sulfur content in coal entering the coking ovens after being mixed by the simultaneous unloading of barges onto a common conveyor. Although the problem may be given an integer programming formulation, the solution method we propose enables one to obtain optimal or near optimal solutions by hand. A hypothetical problem is solved to illustrate the procedure.

Iterative solution of problems of linear programming

Abstract: AN iterative method for solving problems of linear programming has been described in [1]. This method is not completely determinate; its basic cycle is determined by two parameters ϵ and K. Its convergence was proved in [2, 3], a sequence e1 > e2 … > eR … → 0; tending monotonically to zero being used for ϵ; as regards K, the number of iterations in an “internal” procedure B, it was assumed in the proof that the procedure is finalized, i.e. that K = ∞. On the purely theoretical level this internal infinity can easily be avoided (in the same way as when proving the countability of a countable set of countable sets), though to organize practical computations on such a basis would be futile. A “heuristic” method of determining ϵ and K was developed while using the method. The method of selecting ϵ described in [1] required no serious modification. A reliable means of selecting K, based on feedback of the computational data, was only discovered later; it proved to be extremely successful, notably because the major part of the algorithm operating time is spent on the B procedure and a poor choice of K involves a lot of wasted time. In addition, the method was used to solve specific problems of linear programming occurring in the numerical solution of optimal control problems [2, 3]. The particular feature of these problems, namely, that they originate “continuously”, demanded some specialization of the method. The type of specialization and a “qualitative” proof of it will be found in Section 2 below. Section 1 concerns the iterative solution of problems of linear programming with a large number of dimensions. The main expressions underlying the process are the same as in [1], but they are worth repeating here because they make the paper easier to read and take up little space.

Decision making and learning: A simulated marketing manager

Abstract: This paper presents the results of a computerized adaptive learning model. The learner in this case is a hypothetical marketing manager competing with two competitors in a common market. The model is free of heuristic rules-of-thumb and relies instead on a search procedure which updates current decisions based on decisions made and experiences realized in the past. It is adaptable to other learning environments in addition to that given in this paper.

Computational Limitations of Dynamic Programming for Warehouse Location: A Comment

Abstract: practice. Heuristic methods might be used to obtain good, but not necessarily optimal, solutions. Ballou has presented a new dynamic aspect of the warehouse location problem. However, his approach can not handle realistic problems computationally where a company has a reasonable number of warehouses whose locations are interdependent in terms of profitability during a time period. On the other hand, some static solution procedures such as Shycon and Maffei's [2] or Kuehn and Hamburger's [3] can reflect these interdependencies but do not consider the dynamic factors such as demand changes and moving costs. Combination of both approaches in a computationally feasible manner is still in the realm of further research.

Automata-Based Reasoning For Short CircuitDiagnosis In Power Transmission Networks

Abstract: This document introduces an original technique for short circuit diagnosis in power transmission networks. The basic idea is that the diagnostic task can be conveniently viewed as a multi-layer interpretation process of available observations, at the end of which, as a sort of harvest, relevant diagnostic information can be gathered to generate the required diagnosis. Three di erent interpretation layers are considered: local interpretation, global interpretation, and heuristic interpretation. This paper presents the rst (application-independent) two layers. The local interpretation focuses on the behaviour of single protection components that are distributed in the power network and operate when a short circuit occur. The global interpretation provides a global behaviour of the protection apparatus by combining consistent local behaviours so that given interface constraints among components are met. Finally, the heuristic interpretation is meant both to shrink the cardinality of the global interpretation, by eliminating a number of spurious global behaviours on the basis of application-dependent heuristic constraints, and to eventually localize the short circuit and possibly faulty protection components within the transmission network. The implementation of the proposed technique is under way. The resulting system is going to be tested by ENEL, the Italian electricity board, using the transmission network of part of northern Italy. Transactions on Information and Communications Technologies vol 19, © 1997 WIT Press, www.witpress.com, ISSN 1743-3517