Izack Cohen

Managing Stochastic, Finite Capacity, Multi-Project Systems through the Cross-Entropy Methodology

This paper addresses the problem of loading a finite capacity, stochastic (random) and dynamicmulti-project system. The system is controlled by keeping a constant number of projects concurrently in the system. A new approach, based on the Cross-Entropy (CE) method, is proposed to determine optimal loading of the system. Through numerical experiments, we demonstrate the CE method performance and show new insights into its behavior in a noisy system. Particularly, we suggest a trade-off between the convergence time, the number of iterations and the noise level.

Resource allocation in stochastic, finite-capacity, multi-project systems through the cross entropy methodology

Abstract This paper addresses the problem of resource allocation in a finite-capacity, stochastic (random) and dynamic multi-project system. The system is modeled as a queuing network that is controlled by limiting the number of concurrent projects. We propose a Cross Entropy (CE) based approach to determine near-optimal resource allocations to the entities that execute the projects. The performance of the suggested approach is demonstrated through numerical experiments and compared to that of a heuristic, rough-cut based method.

Improving Time-Critical Decision Making in Life-Threatening Situations: Observations and Insights

In this article we present our concept of time-critical decision making, sometimes even in life-threatening situations, and compare it to the process of non-time-critical decision making. Decision-making methodologies have been extensively researched, and some of the published research deals with decision making within the context of everyday life. However, in many organizations it is customary for decisions to be made under pressure and in conditions of uncertainty. Such organizations may benefit from a generic decision-making approach. Two case studies were used to research the characteristics of time-critical decision making. A qualitative analysis of these case studies and previous research insights were integrated. The insights that were found enable us to offer a practical generic approach toward improving the process of time-critical decision making. The suggested approach combines components that were mentioned in previous research with new ones. It contains two phases: The first identifies various decision-making situations in the organization and their classification according to the extent (severity) of time criticality in making and implementing the decision. This classification determines the necessary decision making and implementation procedures, whether they are cognitive or not. The second deals with the relevant components for improving the quality of the decision making. Application of this approach is very simple, and it suits not only military organizations, but also organizations and individuals that will benefit from making better decisions in stressful situations. The approach can be combined with other existing approaches such as risk management.

On the discretized Dubins Traveling Salesman Problem

ABSTRACT This research deals with a variation of the Traveling Salesman Problem in which the cost of a tour, during which a kinematically constrained vehicle visits a set of targets, has to be minimized. We are motivated by situations that include motion planning for unmanned aerial, marine, and ground vehicles, just to name a few possible application outlets. We discretize the original continuous problem and explicitly formulate it as an integer optimization problem. Then we develop a performance bound as a function of the discretization level and the number of targets. The inclusion of a discretization level provides an opportunity to achieve tighter bounds, compared to what has been reported in the literature. We perform a numerical study that quantifies the performance of the suggested approach. The suggested linkage between discretization level, number of targets, and performance may guide discretization-level choices for the solution of motion planning problems. Specifically, theoretical and numerical results indicate that, in many instances, discretization may be set at a low level to strike a balance between computational time and the length of a tour.

Minimizing mortality in a mass casualty event: fluid networks in support of modeling and staffing

The demand for medical treatment of casualties in mass casualty events (MCEs) exceeds resource supply. A key requirement in the management of such tragic but frequent events is thus the efficient allocation of scarce resources. This article develops a mathematical fluid model that captures the operational performance of a hospital during an MCE. The problem is how to allocate the surgeons—the scarcest of resources—between two treatment stations in order to minimize mortality. A focus is placed on casualties in need of immediate care. To this end, optimization problems are developed that are solved by combining theory with numerical analysis. This approach yields structural results that create optimal or near-optimal resource allocation policies. The results give rise to two types of policies, one that prioritizes a single treatment station throughout the MCE and a second policy in which the allocation priority changes. The approach can be implemented when preparing for MCEs and also during their real-time management when future decisions are based on current available information. The results of experiments, based on the outline of real MCEs, demonstrate that the proposed approach provides decision support tools, which are both useful and implementable.

Discretization-Based and Look-Ahead Algorithms for the Dubins Traveling Salesperson Problem

A new class of discretization-based look-ahead algorithms (DLAAs) for the Dubins traveling salesperson problem (DTSP) is presented that compares favorably with the existing algorithms from the literature. The discretization level and the length of the look-ahead horizon are the two parameters that uniquely determine a DLAA, and depending on the application in hand, their values can be easily modified to strike a balance between the execution time and the length of the resulting admissible tour. The time complexity of a DLAA is the sum of two terms, one linear in the number of targets (cities) and one that corresponds to the specification of an initial order for the targets. For instances of the DTSP with densely distributed targets, an algorithm that relies on clustering and leads to shorter tours than the DLAA is also presented.

Kaizen and Stochastic Networks Support the Investigation of Aircraft Failures

Investigating the causes of aircraft failures and preventing their reoccurrence are crucial to achieving and maintaining a high flight safety level; technical failure-analysis teams usually perform these functions. We developed and applied a dual-phased process to improve the investigative procedures that these teams use. In the first phase we used a Kaizen method to reconstruct the investigation process. In the second phase we created a simulation model of the resulting stochastic processing network to evaluate alternative configurations. The results indicated a significant improvement in throughput time and investigation quality. In addition, this unique improvement process could be adapted for use by the many organizations that concurrently run several types of jobs (or projects) in a stochastic and dynamic environment.

On sourcing and stocking policies in a two-echelon, multiple location, repairable parts supply chain

This research develops policies to minimize spare part purchases and repair costs for maintaining a fleet of mission-critical systems that operate from multiple forward (base) locations within a two-echelon repairable supply chain with a central depot. We take a tactical planning perspective to support periodic decisions for spare part purchases and repair sourcing, where the repair capabilities of the various locations are overlapping. We consider three policy classes: a central policy, where all repairs are sourced to a central depot; a local policy, whereby failures are repaired at forward locations; and a mixed policy, where a fraction of the parts is repaired at the bases and the remainder is repaired at the depot. Parts are classified based on their repair cost and lead time. For each part class, we suggest a solution that is based on threshold policies or on the use of a heuristic solution algorithm that extends the industry standard of marginal analysis to determine spare parts positioning by including repair fraction sourcing. A validation study shows that the suggested heuristic performs well compared to an exhaustive search (an average 0.2% difference in cost). An extensive numerical study demonstrates that the algorithm achieves costs which are lower by about 7–12% on average, compared to common, rule-based sourcing policies.