Abstract
The chaotic nature of economical, technological and social environment leads to the need of operational approaches and tools helping managers to detect, to evaluate and to plan what are the required changes for their enterprises. The present work proposes a generic approach which explicitly connects enterprise change methods (Just In Time, Kanban, S.M.E.D., T.P.S., T.Q.M., etc.) and the construction of change trajectories. This approach is based on: enterprise modelling, the definition of a global enterprise performance system, the analysis of change methods, the identification of the initial and the final expected enterprise states, the construction and the evaluation of change trajectories. The systemic framework SAGACE has been enriched and adapted to organise the different modelling languages used to reflect each enterprise modelling view-points. The definition of an enterprise state needs to identify the relevant enterprise performance indicators. These indicators belong to 3 classes: Control performance (Cost, Quality, Delay), Adaptation performance (adaptivity, proactivity), Anticipation performance (flexibility, resources redundancy, standardization, innovation). A set of performance indicators trees reflecting the performance view-points is associated to each enterprise process. Within a process, the Ishikawa approach permits to identify the potential causes of performance alteration. A bottom-up evaluation of the tree nodes based-on the AHP technique valuates the performance indicators. A generic formulation of the enterprise state is proposed. It mainly integrates, for the impacted processes high level performance indicators (Control, Adaptation, Anticipation) allowing to globally observe the impact of change. Moreover it also integrates the operational performance indicators which must be impacted to reach the strategic objectives. They are identified using a like ECOGRAI top-down decomposition technique. An original generic framework integrating life cycle, genericity, and performance dimensions has been proposed to model the change methods. To guide de construction of enterprise trajectories, static and dynamic rules have been identified to determine which change methods can be deployed from a given enterprise state. From the initial state, a trajectory is built by identifying recursively from each state the eligible change methods. The application of a method corresponds to a known performance impact and generates a new enterprise state. The trajectory construction finishes when the objective state is reach or when no more change methods can be employed The proposed approach has been implemented in a tool named Generator of Enterprise Trajectories, using GME© and Matlab© softwares. It has been successfully tested on a manufacturing enterprise using medium scale benchmarks.