On the History of Ecosystem Dynamical Modeling: The Rise and Promises of Qualitative Models.

Ecosystem modeling is a complex and multidisciplinary modeling problem which emerged in the 1950s. It takes advantage of the computational turn in sciences to better understand anthropogenic impacts and improve ecosystem management. For that purpose, ecosystem simulation models based on difference or differential equations were built. These models were relevant for studying dynamical phenomena and still are. However, they face important limitations in data-poor situations. As a response, several formal and non-formal qualitative dynamical modeling approaches were independently developed to overcome some limitations of the existing methods. Qualitative approaches allow studying qualitative dynamics as relevant abstractions of those provided by quantitative models (e.g., response to press perturbations). Each modeling framework can be viewed as a different assemblage of properties (e.g., determinism, stochasticity or synchronous update of variable values) designed to satisfy some scientific objectives. Based on four stated objectives commonly found in complex environmental sciences ((1) grasping qualitative dynamics, (2) making as few assumptions as possible about parameter values, (3) being explanatory and (4) being predictive), our objectives were guided by the wish to model complex and multidisciplinary issues commonly found in ecosystem modeling. We then discussed the relevance of existing modeling approaches and proposed the ecological discrete-event networks (EDEN) modeling framework for this purpose. The EDEN models propose a qualitative, discrete-event, partially synchronous and possibilistic view of ecosystem dynamics. We discussed each of these properties through ecological examples and existing analysis techniques for such models and showed how relevant they are for environmental science studies.

Discrete-event models for the simulation of computed tomography sectors according to hospital structural/organizational changes and expected patient arrival rates.

OBJECTIVE: To analyze the types of computed tomography (CT) scanners most suitable for different hospital sizes and 'scenarios' (exam rates and structural/organizational changes), using discrete-event simulation models. MATERIALS AND METHODS: CT exams were divided into stages, measured during on-site surveys at CT services in small and average size private hospitals. Ten devices in nine health units, five cities and two states of Brazil were studied to this end, and the following data were collected: Time spent in each stage for each type of exam; average monthly number of exams performed and general characteristics of exams. Three arrival rates were defined (103, 154 and 206 patients/day), representing expected demand for the studied units. From these parameters, six scenarios were simulated, consisting of changes in personnel and hospital structure (e.g., 'adding a changing room') in a base scenario (one CT, one changing room, no nursing assistance, arrival rate 1). RESULTS: It was possible to identify a scenario most useful for very large demands, such as large emergency hospitals in big cities, (a CT, nursing assistance and three changing rooms added to the base scenario). Another identified scenario was more adequate for small demands (adding a changing room to the base scenario). CONCLUSION: Administrative/organizational measures are a very important factor in defining productivity in a hospital imaging sector. The focus of these measures should be on detecting bottlenecks and improving processes, regardless of the type of equipment used.