Short term decision support for manufacturing systems is generally difficult because of the initial data needed by the calculations. Previous works suggest the use of a discrete event observer in order to retrieve these data from a virtual copy of the workshop, as up to date as possible at any time. This proposal offered many perspectives, but suffers from the difficulties to generate a decision support tool combining decision calculations and observation. Meanwhile, interesting developments were made in literature about automatic generation of logic control programs for those same manufacturing systems, especially using the Model Driven Engineering. This paper suggests the use of Model Driven Engineering to generate logic control programs, the observer and the decision support tool at the same time, based on the same data collected by the designer of the system. Thus, the last section presents the evolution needed in the initial data structure, as well as the conception flow suggested to automatize the generation.

Non-renewal traffic appears an various shape as the input to nodes an ATM networks. Adequate models of trafic types are required to analyze the performance of ATM switches. Among them are approaches assuming self-similar+ty, TES processes or semi-Markovian processes (SMP). W e study approximative representations of trafic by semi-Markovian processes, which include autocorrelated arrival intervals depending on the states of an underlying Markov chain. Properties of the SMP autocorrelation function are derived, which are helpfil to construct SMP adaptations. ATM swatches with SMP input are analyzed wing eficient algorithms for the discrete time SMP/G/l queue. The method is applied to evaluate the performance of an ATM multiplexer with trafic superposed from a number of ON-OFF sources.

Transitic systems are typically conveying systems used to transport parcels/products. The design of these systems is more and more complex due to flow and speed increase which pose hardware and software problems. The control engineers, who master the logic of control systems, also have to adapt themselves to the type of hardware already present at the customer (PLC) without necessarily mastering their programming. In the current logic of time-tomarket reduction, these designers have to design the control of the systems more and more quickly without denying the Quality. To overcome these challenges, this paper presents a design flow which automates the generation of multiplatform control for the transitic systems. Based on model driven engineering and using a component approach, it has been implemented in a tool. We present the specificities of Straton and Unity generations and analyze the cost of adding a new platform.

International audienceWith the increase of flexibility and production rates, the complexity of manufacturing systems reached a point where the operator in charge of the production activity control of the system is not able to forecast efficiently the impact of his decisions on the global performances. As a matter of fact, more and more Decision Support Systems (DSS) are developed, as much in literature or industrial applications. DSS have one common point: the initialization of their forecasting functionality is based on data coming from the manufacturing system. Furthermore, this feature is fundamental, as it has a direct impact on the accuracy of the forecasts. Considering the variety of input and output data, a data processing is necessary to adapt those coming from the manufacturing system. The aim of this paper is to present several design approaches enabling the integrator of a new manufacturing system to speed up the implementation, with the idea of automate and systematize the maximum design phases thanks the model driven engineering