数据通信中一类信息打包问题的最优解

The Optimal Solutions of a Packing Problem in Data Communication

对于数据通信过程中存在的一类信息打包问题建立了概率模型，并就打包的方案进行讨论，通过定量分析获得了最佳打包方案的存在性，并给出了几个数值模拟结果．

A kind of information packing problem in data communication is dealt with. The information groups arrive according to a point process which is partially finite; the lengths of all groups are identical and independent to the arriving process. For safety and reliability , it is necessary to divide each group into some data pieces, pack and convey them by computer. A probability model is built by which the procedure of the packing can be discussed quantitatively. Stochastic methods are used to determine whether the model has really definite optimal solutions. It is shown that if only suitable total cost in average per time unit and ratio of total cost of additional information to that of useful ones in average per time unit are given, such a model has indeed some intervals as its definite solutions. When the information arrivals are a Poisson process with intensity λ and the length of every data group has exponential distribution with parameter μ -1 , the optimal packing length of information can be found. Some numerical simulations are performed to demonstrate the effectiveness of such models.