Showing papers by "James C. Bezdek published in 1989"

Pool2: a generic system for cognitive map development and decision analysis

Abstract: The authors present Pool2, a generic system for cognitive map development and decision analysis that is based on negative-positive-neutral (NPN) logics and NPN relations. NPN logics and relations are extensions of two-valued crisp logic, crisp (binary) relations, and fuzzy relations, NPN logics and relations assume logic values in the NPN interval (-1, 1) instead of values in (0, 1). A theorem is presented that provides conditions for the existence and uniqueness of heuristic transitive closures of an NPN relation. It is shown that NPN logic and NPN relations can be used directly to model a target world with a combination of NPN relationships of attributes and/or concepts for the purposes of cognitive map understanding, and decision analysis. Two algorithms are presented for heuristic transitive closure computation and for heuristic path searching, respectively. Basic ideas are illustrated by example. A comparison is made between this approach and others. >

Judy Creek: A Case Study for a Two-Dimensional Sediment Deposition Simulation

Abstract: his paper describes a computer program developed at the University of South Carolina to simulate the evolution of carbonate geometries and their facies responding to: (1 ) varying rates of accumulation; (2) eustatic sea-level variation; and (3) tectonic movement of the crust, The simulation creates two-dimensional plots of synchronous depositional sequences within sediment bodies. Rates of carbonate accumulation are modeled as a function of water depth and lateral position across the shelf. Carbonate accumulation includes in situ organic production and transport by hydrodynamic processes. Influx of clastic sediments is modeled to cause an exponential decrease in the rate of carbonate accumulation. Rates of carbonate accumulation can be further diminished with a user-defined depth-controlled wave-damping function. Eustatic variation is modeled by a fourth-order linear change in sea level over time (as described by Vail and others, 1977), with a higher order sinusoidal oscillation of sea level superimposed upon it. Reef margin and interior lagoonal facies on platforms and shelves are predicted. Modeling of these facies zones includes aggradation, progradation, backstepping, shoal development, and drowning. The Devonian Judy Creek reef complex of the western Canada Alberta basin was used both as an aid in constructing the carbonate simulation model and as an example on which to test the completed program. The Judy Creek model was constructed by the fourth co-author (J.C.W.) from the study of 100 cores and the correlation of a systematic grid of wireline log-core cross sections. Judy Creek consists of five overall shoaling-upward depositional cycles. Superimposed upon each cycle are several subcycles. Marine hardgrounds locally cap the first three cycles at the margin of the buildup. The top of the fourth cycle is a subaerial cemented surface. The top of the upper cycle is a widespread marine hardground. Subcycles consist of minor shoaling-upward sequences of lagoonal facies. The results of two versions of the carbonate model include: (1) a site-specific version of the Judy Creek area, which shows the facies location and movement within Judy Creek; and (2) a more generalized carbonate model. Both versions use similar inputs. The geometry and facies distributions of Judy Creek are simulated using a stairlike fourth-order eustatic sea-level curve (as described by Vail and others, 1977) and a low-amplitude higher order sea-level oscillation with a varying period in which sea-level fall is matched by tectonic subsidence. Maximum rates of accumulation average around 3.0 to 5.0 m/10 ka. The fourth-order sea-level variation consists of a rapid 3.5-5.0 m/10 ka rise followed by a gradual rise of 0.7-1.5 m/10 ka, followed by another rapid rise of the same magnitude, followed by another gradual rise. The higher order sea-level oscillation consists of 0.5 to 1.5-m amplitudes with periods ranging from 20 to 40 ka. Tectonic subsidence averages between 0.5 and 1.0 m/10 ka.

An expert system as a component of an integrated system for oil exploration

Abstract: A description is given of the Expert Explorer System, a tool for analyzing areas with the potential for oil production. It has database, pattern recognition, simulation, report generation, and expert system components. The expert system is an essential element of the system because is assists the user in organizing and abstracting data in a manner such that the characteristics of an area under consideration may be compared with the characteristics of other known areas. >

Simulation of Sedimentary Fill of Sichuan Basin by Upper Permian Coals and Carbonate Sequences--An Example of Use of SEDPAK: ABSTRACT

Abstract: The Upper Permian sediments of Sichuan and western Hubei provinces of China, in ascending order, are (1) the basal Ermi basaltic volcanic sequence which exceeds 800-m thickness to the west but is only a few centimeters thick in the east; (2) the overlying coal-bearing Longtan Formation which onlaps as continental coal sequences to the west but downlaps and thins to a few meters of marine shales to the east where it is interbedded with and overlain by thick shelfal carbonates of the Wujiaping Formation; and (3) a thick wedge of carbonates of the Changxing Formation which onlaps the lower coals and interfingers with a thin upper coal sequence to the west. The post-volcanic sedimentary fill can be considered as two supercycles responding to two major sea level changes. The SEDPAK simulation captures deposition of the two supersequences. The first sequence is modeled as onlapping lowstand clastics and coals which, interfingering with ramp carbonates, deepen upward and begin to give up as sea level rise exceeds carbonate growth. The second is modeled as a sequence of ramp to platform carbonates which initially catches up and then keeps up while interfingering westward with coast marine coal sequences. SEDPAK highgrades the hydrocarbon potentialmore » of the carbonate buildups in the second sequence.« less

An Adaptive, Layered Bayes Network

Abstract: An adptive pattern recognition network is described that has several internal feature selection layers. Bayes rule combines features and derives each layer from its predecessor starting from two features per node in the first internal layer. Nodes in higher order layers involve more features than those in the lower order layers. Each node in the last internal layer involves all the input features, and is constructed by different feature combinations. A confidence combination layer then combines recognition confidences of the nodes in the last internal layer. This layer dynamically selects only the most significant (weighted) nodes for each class. Our network provides rapid incremental learning from new training samples, dynamic introduction of new classes and new features, and the exclusion of existing classes and features without retraining on the modified data. We illustrate our method by comparing empirical error rates obtained by applying the layered network, a single internal layer network, and the Bayes quadratic decision rule to the ubiquitous IRIS data.