Showing papers in "Infor in 1988"
TL;DR: An efficient generalized permanent labelling algorithm is presented based on the definition of the concept of a generalized bucket and on a specific order of handling the labels to solve the shortest path problem with time windows.
Abstract: The shortest path problem with time windows (SPPTW) consists of finding the least cost route between a source and a sink in a network G = (N, A) while respecting specified time windows [ai, bi] at each visited node The duration dij of each arc is restricted to positive values while the cost Cij of each arc (i, j) Є A is unrestrictedThis article presents an efficient generalized permanent labelling algorithm to solve this problem This new algorithm is based on the definition of the concept of a generalized bucket and on a specific order of handling the labels The algorithm runs in pseudo-polynomial time Problems with up to 2500 nodes and 250,000 arcs have heen solved
369 citations
TL;DR: In this article, the authors describe trends in knowledge support environments, integrated interactive knowledge acquisition systems and expert system shells, that can provide a specialist community with tools supporting a wide range of knowledge processes, and suggest that such systems provide a major knowledge-based technology with commercial implications and applications going beyond those currently envisioned for expert systems.
Abstract: This paper describes trends in knowledge support environments – integrated interactive knowledge acquisition systems and expert system shells – that can provide a specialist community with tools supporting a wide range of knowledge processes These systems extrapolate the trend from human knowledge engineering, through automated interviewing of the expert, to continuing on-line access to both knowledge acquisition and application processes In knowledge support systems the distinctions between expert, knowledge engineer and client roles are deliberately blurred, and a diversity of knowledge processes and changing roles are supported within an entire interacting community A prototype knowledge support system is described with examples of some ofthe knowledge acquisition and application tools provided It is suggested that such systems provide a major knowledge-based technology with commercial implications and applications going beyond those currently envisioned for expert systems
22 citations
TL;DR: This paper argues that (a) the terminology used in many AI discussions is poor, (b) that many techniques widely touted as revolutionary are ad hoc, “cut and try” methods that will not lead to trustworthy products, and (c) that the fundamental research is more philosophical than practical.
Abstract: It can be said that the most promising field within computer science is Artificial Intelligence, often simply known as AI. Some will interpret this as meaning that AI is a field that holds great pr...
16 citations
TL;DR: This paper explores the use of the augmented Lagrangian in a decomposition method which is a direct application of the “method of multipliers”, and reports some computational evidence from two linear programming problems related to forest management which indicates that empirical convergence rates can be quite satisfactory.
Abstract: In this paper, we explore the use of the augmented Lagrangian in a decomposition method which is a direct application of the “method of multipliers” Contrary to popular opinion, the use of the augmented Lagrangian does not necessarily destroy separability inherent in a problem This separability can be recovered by the use of sequential linearization algorithms, of which the Frank-Wolfe algorithm is an example We outline this decomposition approach in a fairly general framework and then specialize it to the decomposition of linear programs Linear programs are, in at least one sense, the worst case for a method of multipliers approach since little is known about the rate of convergence of the dual multipliers in tMs case We report some computational evidence from two linear programming problems related to forest management which indicates that empirical convergence rates can be quite satisfactory
12 citations
TL;DR: Three algorithms for the directed Steiner tree problem on graphs are presented and their performance compared on a number of problems originally tested by Beasley (1984, 1987) in the case of undirected graphs.
Abstract: : A Steiner Problem in graphs is the problem of finding a set of edges (arcs) with minimum total weight which connects a given set of nodes in an edge- weighted graph (directed or undirected). This paper develops models for the directed Steiner tree problem on graphs. New and old models are examined in terms of their amenability to solution schemes basd on Lagrangian relaxation. As a result, three algorithms are presented and their performance compared on a number of problems originally tested by Beasley (1984, 1987) in the case of undirected graphs. Keywords: Networks, Operations research.
12 citations
TL;DR: In this article, the Dynamic Lot Size Model (DLSM) with stochastic demands is studied, and sufficient conditions for the existence of DH/FH are obtained in terms of salvage functions, state of the system, and decision strategies.
Abstract: In a multiperiod decision problem it is usually the decisions in the first or first few periods that are of immediate importance to the manager. Decision/Forecast horizon (DH/FH) research deals with the question of whether optimal decisions in the first or first few periods (known as DH) can be made without regards to the data from some future period (known as FH) onwards.In this paper the Dynamic Lot Size Model (DLSM) with stochastic demands is studied. The model has a setup cost and hence does not have a convex cost strvicture usually assumed in the earlier DH/FH research involving stochastic problems. It is shown that in the stochastic DLSM, unlike in the convex cost problems, there are examples without DH/FH. However, randomly generated problems seem universally to have decision horizons. Sufficient conditions for the existence of DH/FH are obtained in terms of salvage functions, in terms of states of the system, and in terms of decision strategies. A comparison of these approaches concludes t...
11 citations
10 citations
TL;DR: Comparative simulation test results show that, under varying production conditions, the Lookahead heuristic provides low total inventory and changeover costs while maintaining high service level.
Abstract: This paper presents a simple heuristic for scheduling the production of many items on a single machine when demands are both dynamic and stochastic. The heuristic is based on {s, S) type policies and relies principally on simulations of the upcoming production days (the Lookahead function) to decide if the production facility should be shut down or not on a given working shift. Comparative simulation test results show that, under varying production conditions, the Lookahead heuristic provides low total inventory and changeover costs while maintaining high service level.
9 citations
TL;DR: This paper shows how IRA methods can supplement each other, instead of being viewed as alternatives, and resolves differences between Decision Analysis and Data Analysis by developing different kinds of Prototypes as an integrated framework of IRA.
Abstract: Effective information requirements analysis (IRA) is critical for the success Of application systems. Literature has mainly defined the contingencies under which specific IRA methods are most effective for determining the content of information. This paper shows how IRA methods can supplement each other, instead of being viewed as alternatives. A process for combining IRA methods is developed. Resolving differences between Decision Analysis and Data Analysis by developing different kinds of Prototypes is presented as an integrated framework of IRA. Case studies illustrating this approach are included. The paper extends research in two areas: IRA and Prototyping.
8 citations
TL;DR: The periodic loading problem is formulates, bounds on the solution are given, and it is shown it contains the minimum makespan problem, hence is NPcomplete.
Abstract: The periodic loading problem involves the allocation of periodically-occurring tasks to the time-slots of a synchronous server in such a way as to minimize the maximum utilization of any time-slot. It arises when a time-division-multiplex communications bus collects data from a collection of periodically-producing sensors, e.g., the MIL-STD-1553 protocol. This paper formulates the problem, gives bounds on the solution, and shows it contains the minimum makespan problem, hence is NPcomplete. Heuristic procedures are given and their performance illustrated on a (actual) problem for which a solution is found within 1% accuracy. In addition, an empirical study for the heuristic procedures is conducted on an extensive set of artificial problems.
7 citations
TL;DR: In this article, the authors employ the compatibility matrix approach and derive statistical relationships for the estimation of the mean and standard deviation number of feasible alternsitives plans for a project network consisting of activities that have many alternatives and pairwise compatibility relationships.
Abstract: A project network consists of activities that have many alternatives and pairwise compatibility relationships. This paper employs the compatibility matrix approach and derives statistical relationships for the estimation of the mean and standard deviation number of feasible alternsitives plans.
TL;DR: The degree of interconnection among statements in a program is measured using information theory concepts of entropy and excess-entropy using an interconnectivity table that represents a program via its statements, variables, and control structures.
Abstract: The degree of interconnection among statements in a program is measured using information theory concepts of entropy and excess-entropy. The principal attributes of a program, control flow, data flow and program size, are incorporated in an interconnectivity table that represents a program via its statements, variables, and control structures. The interconnectivity approach is described and some examples are developed to demonstrate its sensitivity. A comparison with current metrics (cyclomatic number and software science) is presented in conjunction with some examples to demonstrate the characteristics of the new metric.
TL;DR: The result is interpreted in function free logic database terms as a transformation of the recursively defined predicate involving: (a) elimination of an argument, and (b) propagation of selections to the extensionally defined predicates.
Abstract: The topic is algebraic optimization of recursive database queries. Queries are expressed by relational algebra expressions including a fixpoint operation. A condition is presented under which a natural join commutes with a fixpoint operation. This condition is a simple check of attribute sets of sub-expressions of the query. The work may be considered a generalization of Aho and Ullman, (1979).The result is interpreted in function free logic database terms as a transformation of the recursively defined predicate involving: (a) elimination of an argument, and (b) propagation of selections (instantiations) to the extensionally defined predicates. A collection of examples shows that this transformation abstracts some optimizations which otherwise are done by more complex graph algorithms (e.g. Bancilhon et al., (1986); Chang, (1985); Gardarin and DeMaindreville, (1986); Henschen and Naqvi, (1984); Kifer and Lozinskii, (1986)). Thus, this optimization is expressed in a form which is not biased towards...
TL;DR: Although Parnas’ criticisms are ill-conceived and poorly aimed, they can be re-con conceived and more appropriately aimed by distinguishing the science of AI from its commercial technology and show that AI has no monopoly in this regard.
Abstract: David Parnas has recently pointed out that artificial intelligence (AI), whatever it is, provides absolutely nothing useful for producing better software faster. His criticisms betray a fundamental misinterpretation of AI and are therefore especially dangerous for naive AI initiates. However, although Parnas’ criticisms are ill-conceived and poorly aimed, they can be re-conceived and more appropriately aimed by distinguishing the science of AI from its commercial technology. The criticisms regain their intended (?) force when recast as complaints about bad science and bad scholarship and, thus re-expressed, show that AI has no monopoly in this regard.To argue that Pamas’ criticisms are ill-conceived requires a commitment about what artificial intelligence is, and why it is useful. Such a commitment will inevitably create more controversy, but should help prevent others from producing equally ill-aimed criticism as a consequence of their ruminations about AI.
TL;DR: Under Poisson arrivals and exponentially distributed service times this paper develops an algorithm for choosing the number of service units and theNumber of waiting spaces so as to minimize the expected total costs per unit time.
Abstract: Service systems are concerned with provision of some form of treatment or accommodation to customers or to something belonging to the customers. Typically, service systems are comprised of two physically distinct areas, one dedicated to the service itself, the other used for customers waiting for service. A limited amount of space is available for the two purposes. This paper addresses a related problem of concern to designers and managing operators of service systems, namely how to divide a total available area among service units and waiting spaces. Under Poisson arrivals and exponentially distributed service times we develop an algorithm for choosing the number of service units and the number of waiting spaces so as to minimize the expected total costs per unit time. Illustrative numerical results are presented.
TL;DR: Specific requirements and criteria for acquisition or construction of software tools suitable in the design phase and problems related to each of two approaches suggested for the development of such tools are identified.
Abstract: Software engineering involves the production and maintenance of high-quality software systems, delivered on time and within budget. To achieve this fundamental goal, software tools constitute essential ingredients because they support methods and languages by automating and systematizing various tasks of the software development process. In this paper, we provide specific requirements and criteria for acquisition or construction of software tools suitable in the design phase. We discuss basic concepts and identify problems related to each of two approaches suggested for the development of such tools.