Jean G. Vaucher

A comparison of simulation event list algorithms

Four algorithms are considered which can be used to schedule events in a general purpose discrete simulation system. Two of the algorithms are new, one is based on an end-order tree structure for event notices, and another uses an indexed linear list. The algorithms are tested with a set of typical stochastic scheduling distributions especially chosen to show the advantages and limitations of the algorithms. The end-order tree algorithm is shown to be an advantageous, immediate replacement for the algorithm in use with current simulation languages. The most promising algorithm uses the indexed list concept. It will require an adaptive routine before it can be employed in general purpose simulators, but its performance is such that further study would be fruitful.

SSJ: SSJ: a framework for stochastic simulation in Java

We introduce SSJ, an organized set of software tools implemented in the Java programming language and offering general-purpose facilities for stochastic simulation programming. It supports the event view, process view, continuous simulation, and arbitrary mixtures of these. Performance, flexibility, and extensibility were key criteria in its design and implementation. We illustrate its use by simple examples and discuss how we dealt with some performance issues in the implementation.

From scenarios to timed automata: building specifications from users requirements

Scenarios as partial behavior description, are used more and more to represent users requirements, and to conduct software engineering. The paper examines automatic generation of specifications from requirements. This is a crucial step when accuracy is desired in the requirement engineering process. Automatic construction of specifications from scenarios reduces to the merging of partial behaviors into global specifications, such that these specifications can reproduce them. The paper presents an incremental algorithm that synthesizes timed automata from scenarios with timing constraints. The algorithm is based on a formalism developed for scenarios. It uses operations semantics, and a mapping between concepts of scenarios, and those of the theory of timed automata.

Pretty-printing of trees

A pleasing layout of printed tree structures is difficult to achieve automatically. The paper points out three main problems. First, the horizontal position of a node on the printed page depends on global consideration of the position of other nodes; secondly, the physical characteristics of printers require scanning the tree in left‐to‐right top‐to‐bottom sequence; finally, page overflow for wide trees must be handled. These problems are illustrated by analysing the shortcomings of a simple printing algorithm. A suitable general binary tree printing algorithm is presented and its adaptation to other types of trees is shown.

Time-dependent shortest paths through a fixed sequence of nodes: application to a travel planning problem

In this paper, we introduce a travel planning problem which is solved by computing time-dependent shortest paths through a fixed sequence of nodes. Given a predetermined itinerary, our travel planning problem consists in finding the best travel plan, involving planes and hotels, based on the travelers preferences. Our time-dependent framework therefore models plane flights, hotels, stays in each city as well as global time constraints. Given the large size of time-dependent networks, an exact decomposition algorithm is devised to solve instances of realistic size in reasonable computation times.

Environmental temperature control in very low birth weight infants (less than 1000 grams) cared for in double-walled incubators

To evaluate the effect of fluctuations in environment and body temperatures on preterm infants, we recorded these variables in very immature newborn infants (birth weight less than 1000 gm) cared for in double-walled incubators (Air-Shields model C-100 and Ohio model IC). Both incubators maintained environmental temperatures corresponding overall to the set point, despite incubator openings. Under skin temperature servocontrol, however, environmental temperature fluctuations were greater than 2 degrees C even in strictly controlled conditions. The pattern of incubator temperature fluctuations depended on the set point rather than on the type of incubator (conventionally heated or heated by warm air blown between the double walls). The long-term clinical significance of the incubator temperature variability remains to be determined; the choice between air and skin servocontrolling should depend in part on the need for environmental stability.

Metabolic and clinical consequences of changing from high-glucose to high-fat regimens in parenterally fed newborn infants*

To evaluate the metabolic and clinical consequences of changing from high-glucose to high-fat regimens during initiation of parenteral nutrition, we performed 22 studies in 11 newborn infants (birth weight (mean +/- SD) 2.54 +/- 0.54 kg, gestational age 37 +/- 3 weeks, postnatal age 8 +/- 3 days) maintained in a constant thermal environment. In a paired design, two isoproteinic (2.4 +/- 0.2 gm/kg/day) and isocaloric (64 +/- 6 kcal/kg/day) regimens differing by source of energy (high glucose vs high lipid) were infused on consecutive days. Environmental and body temperatures were recorded during a 4-hour period, and 24-hour urinary excretions of catecholamines, nitrogen, and C peptide were measured. Despite constant incubator and average skin temperatures, the rectal and interscapular temperatures rose significantly when the high-glucose regimen was changed to a high-lipid regimen. The specific locations of these changes in body temperature suggested brown fat activation. A significant drop in nitrogen retention (63 +/- 9% vs 56 +/- 10%) during the lipid infusion could be further evidence of a metabolic adaptation to the rapid change in energy substrates.

Discovering New Change Patterns in Object-Oriented Systems

Modern software has to evolve to meet the needs of stakeholders; but the nature and scope of this evolution is difficult to anticipate and manage. In this paper, we examine techniques which can discover interesting patterns of evolution in large object-oriented systems. To locate patterns, we use clustering to group together classes which change in the same manner at the same time. Then, we use dynamic time warping to find if a group of classes is similar to another when we ignore the exact moment when changes occur. Groups that exhibit distinctive evolution properties are potential candidates for new evolution patterns. Finally, in a study of two industrial open-source libraries, we identified four new types of change patterns whose usefulness is determined by perusal of the release notes and the architecture.

Experimenting with Gnutella Communities

Computer networks or distributed systems in general may be regarded as communities where the individual components, be they entire systems, application software or users, interact in a shared environment. Such communities dynamically evolve with components or nodes joining and leaving the system. Their own individual activities affect the communitys behavior and vice versa. This paper discusses various practical experiments undertaken to investigate the behavior of a real system, the Gnutella network, which represents such a community. Gnutella is a distributed Peer-to-Peer data-sharing system without any central control. It turns out that most interactions between nodes do not last long and much of their activity is devoted to finding appropriate partners in the network. The experimental results presented have been obtained from a Java implementation of Gnutella running in the open Internet environment, and thus in unknown and quickly changing network structures heavily depending on chance.

On the Distribution of Event Times for the notices in a Simulation Event List

AbstractThe time required to insert event notices in a simulation event list has a critical effect on the speed of execution of simulation programs. Even though several event list data structures and algorithms have been proposed to reduce this overhead, these attempts have neglected the theoretical aspects of the problem ; and no analysis has ever been made for even the simplest event list algorithms. In this paper, renewal theory is used to obtain the steady-state distribution for the scheduled event times. With the results, one can compute the performance of the standard implementation of the event list algorithm. More significantly, the results provide insight into the operation of event list algorithms and suggest strategies for their improvement. Also given, is an approximate expression for the duration of the transition period preceding steady state.