The learning rate for the distributed simulation in cache design in operating system

We propose a cache of a structure with two classes. The paper presents the solution to the function of the first class, i.e., allocating. Allocating the cache for all tasks in real-time is a new approach. This makes it more feasible for OS (operating system) activity. The far node with long transmission time and the task rarely repeated are all served without getting the time out message or slowing down the overall system performance. To accomplish this we aim for the steepest rate for moving to the optimum point of the implied function of allocation. What is worth mentioning is that by combining time series methods with the elements of orthogonal, reduces the chaos and lets analytical methods of variational calculus do the job, because the tasks more or less possess characteristics of inertia