John A. Sauter

Changing Roles: Dynamic Role Assignment

Modeling roles as sets of normative behavior that agents can assume has been found to be a useful development technique. An important characteristic of real-world agent systems is that the roles played by an agent may change over time. These changes can be of several different kinds. We describe an illustrative application where such role changes are important, analyze and classify the various kinds of role changes over time that may occur, and show how this analysis is useful in developing a more formal description of the application.

Oxygen requirement for prostaglandin biosynthesis.

The formation of prostaglandins by vesicular gland cyclooxygenase can be regulated by O2 concentrations below 30 micron (20 mm Hg) with an apparent Km value of about 5 micron. This result suggests that most mammalian tissues might be expected normally to have sufficient O2 for some synthesis, and that only occasionally, would the level of oxygen available within some tissues become marginally inadequate to sustain a high rate of prostaglandin biosynthesis.

Swarming Unmanned Air and Ground Systems for Surveillance and Base Protection

The emergence of various risks to global security and stability is a motivation to develop remote sensing and monitoring systems that can be deployed on Unmanned Vehicles (UxVs). This requires the development of robust autonomous control technologies that can reliably coordinate large numbers of networked heterogeneous systems cooperating on a common mission objective. This paper describes a promising approach to addressing this challenge by using swarm intelligence to coordinate multiple heterogeneous vehicles and remote sensors in realistic applications. We describe a class of stigmergic algorithms based on digital pheromones to control and coordinate the actions of heterogeneous unmanned air and ground systems in two applications: broad area surveillance and base protection. An Operator System Interface was developed to evaluate techniques for enabling a single operator to monitor and manage multiple unmanned vehicles and unattended sensors of different types. The results from recent demonstrations of the technology using air and ground platforms are reported.

A systematic market approach to distributed constraint problems

MarCon (Market-based Constraints) applies market-based control to distributed constraint problems. It offers a new approach to distributing constraint problems that avoids challenges to current approaches in some domains, and it provides a systematic way to apply markets to many problems. Constraint agents interact with one another via the variable agents in which they share an interest, expressing their preferences over sets of assignments. Each variable integrates this information from the constraints interested in it and provides feedback that enables the constraints to shrink their sets of assignments until they converge on a solution. MarCon has been tested in the domain of mechanical design, in which its set-narrowing process is particularly useful.

Effectiveness of Digital Pheromones Controlling Swarming Vehicles in Military Scenarios

The use of digital pheromones for controlling and coordinating swarms of unmanned vehicles has been studied under various conditions. This paper describes experiments that demonstrate their effectiveness in several militarily significant scenarios in simulation. The scenarios are described along with the results comparing the performance of swarming to traditional military tactics. A demonstration was conducted using these same algorithms to coordinate unmanned vehicles in a simulated exercise. Two air vehicles controlled by digital pheromones and four ground robots controlled by a related stigmergic algorithm successfully executed a two-hour multi-mission surveillance, patrol, target acquisition, and tracking scenario. The vehicles were given only high-level instructions, such as “survey this areaandidentifyandtrackanytargets”or“patrolaroundthisconvoy”.Theairvehicleswere abletodynamicallyadapttonewcommandsandcoordinatetheiractionswitheachotherand the ground robots to achieve the objectives. The algorithm’s robustness was demonstrated when it dynamically adjusted to the unplanned failure of one of the ground robots without any operator intervention.

Selective effects of fatty acids upon cell growth and metabolic regulation.

Positional isomers ofcis-methyleneoctadecanoic acid differed greatly in their efficiency for growth of an unsaturated fatty acid auxotroph ofEscherichia coli upon glucose as a carbon source. The 8, 9, and 11 isomers were more efficient in producing cells (60–70 cells/fmole) than the others (0–7 cells/fmole), although all isomers were found esterified to a similar extent into cellular lipid. WithSaccharomyces cerevisiae mutants, all isomers between 6 and 12 supported some growth of the eukaryotic cells, and the 7 and 9 isomers were slightly more efficient than the 8-isomer. WhenE. coli were grown with glycerol, all isomers from 5 to 14 supported growth, and those with the substituent near the center of the acyl chain had the greatest efficiency (70 cells/fmole). With the glycerol medium, the pattern of efficiencies for the variouscis-methylene acyl chains resembled the broad selectivity reported earlier for thecis-ethylenic isomers in glucose medium, which agreed closely with predictions based upon the physical property of their phospholipid derivatives. Thus, metabolism of glycerol appeared to allow the cyclopropane acyl chains to support cell functions to the limits expected for bulk phase chain-chain fluidity considerations. This broad specificity was also obtained when cells were grown on glucose with cyclic AMP added to the culture. Therefore, the selective inadequacies of the 5, 6, 7, 10, 12 and 13 isomers in supporting cell growth on glucose may occur through an interaction modified by cAMP and dependent upon reduced cellular levels of cyclic AMP. The highly selective pattern of efficiency of thecis-methylene acids forE. coli growth on glucose resembles that with the acetylenic acids, but was shifted one carbon atom toward the methyl terminus. This observed selectivity pattern seems due to interactions of the individual acyl chains with cellular protein(s) rather than to chain-chain interactions in a bulk phase. The ability of certain positional isomers to support cell function equally well in both nutrient conditions suggests that the role of those acyl chain isomers may be independent of metabolite flux or cyclic nucleotide contents of the cell, whereas the actions of other isomeric fatty acids seem closely related to the metabolic status of the cell. A highly selective role for different fatty acids in modulating cellular function seems possible on the basis of the current evidence.

Swarming methods for geospatial reasoning

Geospatial data are often used to predict or recommend movements of robots, people, or animals (‘walkers’). Analysis of such systems can be combinatorially explosive. Each decision that a walker makes generates a new set of possible future decisions, and the tree of possible futures grows exponentially. Complete enumeration of alternatives is out of the question. One approach that we have found promising is to instantiate a large population of simple computer agents that explore possible paths through the landscape. The aggregate behaviour of this swarm of agents estimates the likely behaviour of the real‐world system. This paper will discuss techniques that we have found useful in swarming geospatial reasoning, illustrate their behaviour in specific cases, compare them with existing techniques for path planning, and discuss the application of such systems.

Demonstration of Digital Pheromone Swarming Control of Multiple Unmanned Air Vehicles

The use of digital pheromones for controlling and coordinating swarms of unmanned vehicles has been studied under various conditions demonstrating their effectiveness in multiple military scenarios in simulation. An experiment was conducted to verify that these same algorithms could effectively coordinate unmanned vehicles in a simulated exercise. Two air vehicles (modified target drones) controlled by digital pheromones and four ground robots controlled by a related stigmergic algorithm successfully executed a two-hour multi-mission surveillance, patrol, target acquisition, and tracking scenario without any scripting. The vehicles were given only high-level instructions, such as “survey this area and identify and track any targets” or “patrol around this convoy”. The air vehicles were able to dynamically adapt to new commands and coordinate their actions with each other and the ground robots to achieve the objectives. The algorithm’s robustness was demonstrated when it dynamically adjusted to the unplanned failure of one of the ground robots without any operator intervention.

Toward the Specification and Design of Industrial Synthetic Ecosystems

Many agent-based systems rely for their effectiveness on the intelligence of individual agents, and interaction among agents is required simply to coordinate these individually complex decisions. Specification and design methods for such systems focus on the internal architecture of individual agents. An alternative approach, “Synthetic Ecosystems,” uses relatively simple agents and draws heavily on the dynamics of the interaction among these agents as well as their internal processing to solve domain problems. The specification and design of such systems must include not only the individual agents, but also the structure and dynamics of their interaction. This paper briefly defines and motivates the Synthetic Ecosystems approach and outlines some techniques that have proven useful in specifying and designing them.

Multi-layer simulation for analyzing IED threats

IEDs, made infamous in Iraq and Afghanistan, are a potential terrorist weapon anywhere in the world. Most counter-IED technologies are targeted at the point of the blast. DEFUSE can interdict and disrupt OPFOR activities left of the blast, well before planting and detonating the IED. The system integrates three layers of simulation: social, process, and environmental.