Queue Control Method Based On Marginal Effect Global Optimal

As we all know, process control methods play a very important role in a network node queue for data transmission scheduling of resource allocation types, and resolving conflicts congestion. They follow queue control method setting various data streams to the switching node to allocate bandwidth resources and services, so that all the data streams can enter at their own priorities through node. According to different service rules and queue control method, there are many kinds of existing access control methods of queue management mode wireless communication. Overall, they can be classified by tending to favor class or fairness.Methods tending to fairness, such as DropTail, Random Drop, Drop Front and so on.This type of algorithms will take the needs of all the transmission queue formed by the data streams into account. And the result is that the important information streams cannot always pass through the node quickly.Methods tending to favor class, such as CBQ(Class Based Queue) and so on. They are basically built on the basis of fairness algorithms and adding priority factors. These algorithms guarantee the QOS(Quality of Service), but have Sacrificed fairness between the data streams When the information reaches the transmission queue with high priority, in these algorithms the nodes sent the information with high priority first. Bandwidth was completely diverted by high-priority data stream. At the same time most of them cannot get rid of the weaknesses of the fairness algorithms they are based on.What presented in this article is a queue management mechanism based on the marginal effect of economics and game modes. Fairness and priority will be taking into account in this mechanism, demanding for bandwidth in different queues after a game to achieve optimal under conditions of a proposed measure. The optimal solution is derived ultimately affected the marginal effect parameters.Marginal utility, sometimes referred to as the marginal contribution, refers to when the consumer successive increase per consumer unit,unit revenue is brought gradually decreasing (Although bringing the total revenue is still increasing.).In this paper, the marginal effect exists between different business users’ satisfaction and the bandwidth provided by the nodes. In another word, in the process of increasing bandwidth provided the node, although the satisfaction of customer is increasing,but satisfaction brought by unit bandwidth is decreasing.While the parameters is in reasonable range, in the process of increasing high priority data flow’s bandwidth, the latter part of the reduced post-unit revenue will slowdown the increasing of total income. Meanwhile the low-priority data stream may make the overall revenue growth rate batter than the high-priority stream. At this time, allocating part if the bandwidth to low-priority stream is reasonable.