یک روش چند عاملی جدید مبتنی بر یادگیری تقویتی برای شکل‌دهی ترافیک و تخصیص حافظه بافر در روترها

Normal 0 false false false EN-US X-NONE AR-SA MicrosoftInternetExplorer4 /* Style Definitions */ table.MsoNormalTable {mso-style-name:"Table Normal" mso-tstyle-rowband-size:0 mso-tstyle-colband-size:0 mso-style-noshow:yes mso-style-priority:99 mso-style-qformat:yes mso-style-parent:"" mso-padding-alt:0cm 5.4pt 0cm 5.4pt mso-para-margin:0cm mso-para-margin-bottom:.0001pt mso-pagination:widow-orphan font-size:11.0pt font-family:"Calibri","sans-serif" mso-ascii-font-family:Calibri mso-ascii-theme-font:minor-latin mso-fareast-font-family:"Times New Roman" mso-fareast-theme-font:minor-fareast mso-hansi-font-family:Calibri mso-hansi-theme-font:minor-latin mso-bidi-font-family:Arial mso-bidi-theme-font:minor-bidi} Abstract In this paper, realizing the distributed structure of computer networks, the random behaviors in such networks, and the time limitations for control algorithms, the concepts of reinforcement learning and multi-agent systems are invoked for traffic shaping and buffer allocation between various ports of a router. In fact, a new traffic shaper based on token bucket has been developed. In this traffic shaper, instead of a static token production rate, a dynamic and intelligent rate based on the network condition is specified. This results in a reasonable utilization of bandwidth while preventing traffic overload in other part of the network. Besides, based on the stated techniques a new method for dynamic buffer allocation in the ports of a router is developed. This leads to a reduction in the total number of packet dropping in the whole network. Simulation results show the effectiveness of the proposed techniques.