81 
A distributed graph reducer for lazy functional languages /Howson, Christopher January 1992 (has links)
This thesis describes a model for distributed graph reduction implemented on a network of transputers. The model allows variable size communications between processors, by exchanging subgraphs instead of single nodes. Functional languages with lazy semantics have graph nodes representing unevaluated arguments. These nodes require special treatment by the run time system because they must not be copied. By checking the structure of the transmitted subgraphs, it is possible to determine which unevaluated expressions have no external references to them and so may safely be included in the subgraph with no overhead. This allows large subgraphs to be exchanged while reducing the demands on the communications system. This technique raises the possibility of implementations on a wide variety of distributed computers such as networks of workstations, which hitherto has been considered impractical.

82 
A partial implementation of the contour model of block structured processesLaviana, Lyn D January 1974 (has links)
No description available.

83 
Enhancing network communication in NPSNETV virtual Environments using XML : Described dynamic Behavior (DBP) ProtocolsFischer, William D. 09 1900 (has links)
The existing component protocols, as well as new protocols introduced at runtime into NPSNETV are written in their native programming language. As a result, they require authoring and compiling by a trained programmer. The long time frame required to change or introduce new protocols into NPSNETV, a dynamically extensible virtual environment, detracts from the dynamicism of the virtual environment. Networking optimization thresholds to support NPSNETV needed to be determined to ensure that the networking is performed efficiently, and system resources to other systems, such as graphics rendering, are maximized. This thesis implements component protocols described using Extensible Markup Language (XML) into NPSNETV. These protocols are created with different fidelity resolutions for each protocol, which can be swapped at runtime based on the network state. Network testing was performed to find the ideal maximum packet rates based on the impact on CPU utilization and packet loss. By using XML, nonprogrammers can edit protocols for inclusion in a simulation at runtime. Important contributions include adding protocols to NPSNETV with highresolution and lowresolution versions, described by XML documents. Basic network optimization is added to NPSNETV to take advantage of the protocolsÎ± resolution switching ability. The network testing revealed a linear correlation between the packet sending rate and CPU utilization, and a polynomial correlation between the packet sending rate and percentage packet loss. / US Army (USA) author

84 
Determination of reservoir daily operation policies by stochastic dynamic programmingTsou, C. Anthony January 1970 (has links)
Reservoir operation policies are often formulated deterministically on the basis of critical flow hydrology. However, if a dynamic river daily flow forecast system is available for the whole season, the forecast information should be fully utilized in reservoir regulation. Given such a forecast system, two approaches to determining optimal daily operation policies for a single purpose flood control reservoir are suggested. Both approaches use stochastic dynamic programming: one involves the minimization of the expected value of flood damages, and the other involves minimizing the probability of occurrence of an undesirable event, which is a flood damage exceeding a certain amount. The probabilistic approach not only offers a set of alternative optimal daily operation policies, but also indicates the probabilities of being able to achieve the objectives, and thus it forms a basis for comparing and evaluating the alternative objectives. / Applied Science, Faculty of / Civil Engineering, Department of / Graduate

85 
Duality in convex programmingMuir, David Charles William January 1966 (has links)
Problems of minimizing a convex function or maximizing a concave function over a convex set are called convex programming problems. Duality principles relate two problems, one a minimization problem, the other a maximization problem, in such a way that a solution to one implies a solution to the other and that the minimum value of one is equal to the maximum value of the other.
When the functions are linear and the constraint sets are polyhedral, the problems are called linear programming problems. Their duality is wellknown. Certain duality results of linear programming can be extended to convex programming by means of the theory of conjugate convex functions introduced by Fenchel ([1], [2]).
In this thesis the theory of conjugate functions is generalized and applied to convex programming problems. In particular a duality theorem is given for a class of convex programming problems. This theorem is compared with a duality theorem for convex programming problems given by Dorn [3]. / Science, Faculty of / Mathematics, Department of / Graduate

86 
A convex hull algorithm optimal for point sets in even dimensionsSeidel, Raimund January 1981 (has links)
Finding the convex hull of a finite set of points is important not only for practical applications but also for theoretical reasons: a number of geometrical problems, such as constructing Voronoi diagrams or intersecting hyperspheres, can be reduced to the convex hull problem, and a fast convex hull algorithm yields fast algorithms for these other problems.
This thesis deals with the problem of constructing the convex hull of a finite point set in R . Mathematical properties of convex hulls are developed, in particular, their facial structure, their representation, bounds on the number of faces, and the concept of duality. The main result of this thesis is an O(nlogn + n[(d+1)/2]) algorithm for the construction of the convex hull of n points in Rd. It is shown that this algorithm is worst case optimal for even d≥2. / Science, Faculty of / Computer Science, Department of / Graduate

87 
Multicriteria linear fractional programmingChoo, EngUng January 1981 (has links)
The object of this thesis is to study the multicriteria linear fractional programming problems (MLFP).
The characterizations of efficiency, weak efficiency and proper efficiency are derived. In the bicriteria case, the set E of all efficient solutions of (MLFP) is pathconnected by a finite number of line segments and the efficient frontier F(E) can be evaluated by using the row parametric technique in linear programming. The weakly efficient (respectively, properly efficient) solutions can be generated by solving the generalized Tchebycheff norm problems (Tβ) (respectively, (βα)) with different parameters β and α. The set E[sup=W] of all weakly efficient solutions of (MLFP) is compact and pathconnected by finitely many line segments.
A dual problem is formulated which is a natural extension of the usual dual in linear programming, the Wolfe dual in nonlinear programming and the Isermann dual in multiple objective linear programming. Duality results are established under the assumption that the criteria are concave functions. The matrix of the dual variables of (MLFP) can be evaluated by solving L linear programs.
A heuristic arrow search algorithm is developed for solving general multicriteria programming problems interactively. The decision maker merely selects his most preferred one amongstthe presented alternatives. Solutions generated are evenly distributed over the desired neighbourhood
of the weakly efficient frontier. The algorithm is convergent in the bicriteria case, with appropriate convexity conditions. When applied to solve (MLFP), the arrow search algorithm uses only the linear programming techniques. / Business, Sauder School of / Graduate

88 
Program manipulation using a grammarbased metaprogramming systemCameron, Robert Douglas January 1983 (has links)
Program manipulation is defined as the automated or semiautomated tailoring, modification or transformation of software to achieve programs with improved performance or other characteristics or to adapt programs to alternative uses. Program manipulation programs (and other sorts of software tool as well) are characterized as "programs about programs" or metaprograms. It is proposed that the implementation of these metaprograms be aided by a new type of metatool, the metaprogramming system, which is characterized by its systematic treatment of programs as data objects.
In support of the metaprogramming system concept, a grammarbased scheme (called GRAMPS) for specifying metaprogramming systems is put forward. This scheme is illustrated in its particular application to the programming language Pascal (yielding a specification for the metaprogramming system Pascal MPS). The suitability of GRAMPSbased systems for implementing metaprograms is evaluated through theoretical consideration of the properties of GRAMPS systems in general, and an analysis of programming experiments using Pascal MPS in particular.
The use of metaprogramming systems for program manipulation tasks is analysed with respect to several application areas. Metaprogramming systems are compared with various alternatives for the implementation of program transformations. Several types of programming utility are discussed including the use of metaprograms to develop versions of target programs specialized for program analysis. The use of metaprogramming systems in the implementation of syntactic editors and integrated program manipulation systems is also considered.
The results presented support both the GRAMPS scheme itself and the grammarbased treatment of programs as data objects in general. From a theoretical viewpoint, GRAMPS was found to have four important completeness properties. From an experimental perspective, Pascal MPS was found to be generally convenient and practical for largescale applications. In each of the application areas considered, the metaprogramming system approach compared favorably with available alternatives. / Applied Science, Faculty of / Electrical and Computer Engineering, Department of / Graduate

89 
The role of exception mechanisms in software systems designAtkins, Margaret Stella January 1985 (has links)
Exception handling is a crucial aspect of practical programming, particularly in systems allowing logical concurrency such as multiprocess distributed systems.
First, a survey of existing exception handling mechanisms in operating systems is performed, which shows a diversity of implementations, depending on the process model and the method of interprocess communication. The thesis then develops a model for designing software which exploits the different mechanisms for handling normal and exceptional events. The model is applicable in many multiprocess programming environments, and not only preserves modularity, but also enhances efficiency and reliability, while often increasing concurrency.
To derive such a model, exceptions in multiprocess software are classified primarily according to the program level at which they are detected and handled. Servertoclient exceptions are of particular interest because of their ubiquity; these are exceptions detected by a server and handled by a client.
The model treats systems programs as event driven, and proposes dividing the events into normal or exceptional, according to the cost and mechanisms for handling them. Techniques are described for designing software according to three criteria: minimising the average runtime, minimising the exception processing time, and incrementally increasing the program's functionality.
Many examples are given which illustrate the use of the general model.
Program paradigms in several languages and in several systems are introduced to model features which are system dependent, through illustrative examples for asynchronous i/o multiplexing, and for exception notification from a server to its client or clients. Finally, some programs which have been implemented according to the rules of the model are described and compared with their more conventional counterparts. These programs illustrate the practicality and usefulness of the model for diverse systems and concurrent environments. / Science, Faculty of / Computer Science, Department of / Graduate

90 
Necessary conditions for a solution of a nonlinear programming problemLee, Linda May January 1973 (has links)
The conditions required for a solution of general nonlinear programming problems of the form
min{f(x): x є X, g(x) ≤ 0, h(x)=0};
where f is called the objective function, g the inequality constraint and. h the equality constraint, are presented in this thesis. The following cases are studied:
(1) X, a finite dimensional space; f, a real valued function; and g and h finite dimensional vector functions.
(2) X, an infinite dimensional space; f, a real valued function; and g and h either finite or infinite dimensional vector functions.
An application of this type of problem to optimal control will be given and the recent developments in this area will be discussed. / Science, Faculty of / Mathematics, Department of / Graduate

Page generated in 0.167 seconds