Abhijit Deshmukh

Complexity in manufacturing systems, Part 1: Analysis of static complexity

This paper studies static complexity in manufacturing systems. We enumerate factors influencing static complexity, and define a static complexity measure in terms of the processing requirements of parts to be produced and machine capabilities. The measure suggested for static complexity in manufacturing systems needs only the information available from production orders and process plans. The variation in static complexity is studied with respect to part similarity, system size, and product design changes. Finally, we present relationships between the static complexity measure and system performance.

Analysis of supply contracts with quantity flexibility

This paper explores a class of supply contracts under which a buyer receives discounts for committing to purchases in advance. The further in advance the commitment is made, the larger the discount. As time rolls forward, the buyer can increase the order quantities for future periods of the rolling horizon based on updated demand forecast information and inventory status. However, the buyer pays a higher per-unit cost for the incremental units. Such contracts are used by automobile and contract manufacturers, and are quite common in fuel oil and natural gas delivery markets. We develop a finite-horizon dynamic programming model to characterize the structure of the optimal replenishment strategy for the buyer. We present heuristic approaches to calculate the order volume in each period of the rolling horizon. Finally, we numerically evaluate the heuristic approaches and draw some managerial insights based on the findings.

Performance prediction of an unmanned airborne vehicle multi-agent system

Consider unmanned airborne vehicle (UAV) control agents in a dynamic multi-agent system. The agents must have a set of goals such as destination airport and intermediate positions. At the same time, the agents have to avoid gun shootings which move to their neighbors randomly. Agents try to build and execute plans that yield a high probability of successfully achieving the targets. The plans are developed based on the negotiations between different UAVs in the region with the overall goal in mind. The information about enemy defenses can be communicated between UAVs and they can negotiate about the paths to be taken based on their resources, such as fuel, load, available time to complete the task and the information about the threat. By constructing a Markov Decision Process (MDP) in this paper, we derive the optimal path. Combining the MDP and the sample path technique, we obtain the maximum probability that the UAVs successfully reach the target.

A hybrid approach to manufacturing enterprise simulation

Manufacturing enterprise decisions can be classified into four groups: business decisions, design decisions, engineering decisions, and production decisions. Numerous physical and software simulation techniques have been used to evaluate specific decisions by predicting their impact on the system as measured by one or more performance measures. We focus on production decisions, where discrete-event simulation models perform that evaluation. We argue that such an evaluation is limited in time and scope, and does not capture the potential impact of these decisions on the whole enterprise. We propose integrating these discrete-event models with system dynamic models and we show the potential benefits of such an integration using an example of semiconductor enterprise.

Multiscale decision-making: Bridging organizational scales in systems with distributed decision-makers

Decision-making in organizations is complex due to interdependencies among decision-makers (agents) within and across organizational hierarchies. We propose a multiscale decision-making model that captures and analyzes multiscale agent interactions in large, distributed decision-making systems. In general, multiscale systems exhibit phenomena that are coupled through various temporal, spatial and organizational scales. Our model focuses on the organizational scale and provides analytic, closed-form solutions which enable agents across all organizational scales to select a best course of action. By setting an optimal intensity level for agent interactions, an organizational designer can align the choices of self-interested agents with the overall goals of the organization. Moreover, our results demonstrate when local and aggregate information exchange is sufficient for system-wide optimal decision-making. We motivate the model and illustrate its capabilities using a manufacturing enterprise example.

Unifying temporal and organizational scales in multiscale decision-making

In enterprise systems, making decisions is a complex task for agents at all levels of the organizational hierarchy. To calculate an optimal course of action, an agent has to include uncertainties and the anticipated decisions of other agents, recognizing that they also engage in a stochastic, game-theoretic reasoning process. Furthermore, higher-level agents seek to align the interests of their subordinates by providing incentives. Incentive-giving and receiving agents need to include the effect of the incentive on their payoffs in the optimal strategy calculations. In this paper, we present a multiscale decision-making model that accounts for uncertainties and organizational interdependencies over time. Multiscale decision-making combines stochastic games with hierarchical Markov decision processes to model and solve multi-organizational-scale and multi-time-scale problems. This is the first model that unifies the organizational and temporal scales and can solve a 3-agent, 3-period problem. Solutions can be derived as analytic equations with low computational effort. We apply the model to a service enterprise challenge that illustrates the applicability and relevance of the model. This paper makes an important contribution to the foundation of multiscale decision theory and represents a key step towards solving the general X-agent, T-period problem.

Characteristics of part mix complexity measure for manufacturing systems

The authors provide a general framework for measuring planning and scheduling complexity associated with a group of parts to be manufactured simultaneously. An application of this research could be in generating part mixes with a wide range of part mix complexity for simulation models of manufacturing systems, so as to test the system performance under different levels of complexity. The authors identify special properties of the proposed information-theoretic entropy measure that can be used by system planners to select a set of parts or operations from given alternatives. They study the behavior of this complexity measure with respect to part similarity, and show synergistic effects of different variables on this measure.<<ETX>>

A Bayesian network approach to energy-aware distributed sensing

This paper presents a strategy for the design and implementation of an energy-efficient multi-sensor network, based on the structure of sectioned Bayesian networks. A key issue in the design of Bayesian networks for monitoring engineering systems is to ensure that a reliable inference scheme about the health of the system can be made by combining information acquired from each sensor in the system into a single Bayesian network. However, as the size of the network rapidly grows, aggregating information made by all the sensors becomes computationally intractable. Hence, sectioning of the Bayesian network based on functional or logical constraints allows for improved computational efficiency in aggregating information while reducing the overall communication requirements. This ultimately leads to a reduction of the energy cost which is critical to effective operation of the sensor network.

Fundamental Research into the Design of Large-Scale Complex Systems

In February, 2010, the National Science Foundation sponsored a workshop that brought together forty-seven representatives from the US Department of Defense, US Air Force, NASA, the National Institute for Standards and Technology, the National Science Foundation, and academic researchers from systems engineering, engineering design, psychology and cognition, organization design and innovation, and economics and mathematics to construct a research program into foundational issues that underlie how we design and develop large-scale complex systems such as aircraft, spacecraft, and launch systems. This paper is the first report of the results of that conference. This abstract is being submitted only thirty-six hours after the workshop, so that the results have not yet been digested and reviewed. However, the abstract will cover the intent and plan of the workshop.

Hierarchically Organized Bayesian Networks for Distributed Sensor Networks

As sensor hardware becomes more sophisticated, smaller in size and increasingly affordable, use of large scale sensor networks is bound to become a reality in several application domains, such as vehicle condition monitoring, environmental sensing and security assessment. The ability to incorporate communication and decision capabilities in individual or groups of sensors, opens new opportunities for distributed sensor networks to monitor complex engineering systems. In such large scale sensor networks, the ability to integrate observations or inferences made by distributed sensors into a single hypothesis about the state of the system is critical. This paper addresses the sensor integration issue in hierarchically organized sensor networks. We propose a multi-agent architecture for distributed sensor networks. We present a new formalism to represent causal relations and prior beliefs of hierarchies of sensors, called Hierarchically Organized Bayesian Networks (HOBN), which is a semantic extension of Multiply Sectioned Bayesian Networks (MSBN). This formalism allows a sensor to reason about the integrity of a sensed signal or the integrity of neighboring sensors. Furthermore, we can also evaluate the consistency of local observations with respect to the knowledge of the system gathered up to that point.Copyright