Аннотация:Until recently, the standard communication model for scheduling the task of a parallel program has been the homogeneous communication model (also known as the delay model) introduced by Rayward-Smith for unit-execution-times, unit-communication times (UET-UTC) precedence graphs. In this model, we have a set of identical processors that are able to communicate in a uniform way. We want to use these processors in order to process a set of tasks that are subject to precedence contraints. Each task has a processing time, and if two adjacent task of the precedence graph are processed by two different processors (resp. the same processors) then a communication delay has to be taken into account explicitly (resp. the communication time is neglected). The problem is to find a trade-off between the two extreme solutions, namely, execute all the tasks sequentially without communications, or try to use all the potential parallelism but in the cost of an increased communication overhead. This model has been extensively studied these last years both from the compexity and the (non)-approximability point of views.