在基於傳統的蜂巢式(Cellular)無線通訊系統中,裝置間通訊(Device to Device Communication)能使D2D Pairs之間不需透過基地台(eNB)來進行通訊,讓D2D Pairs與eNB之間的跳躍增益降低來減低eNB所需的負擔,及使用有效的資源管理來分配每一對D2D Pairs所使用的資源。透過以上幾點來提高無線資源頻譜的效率。本論文主要探討在LTE系統架構下,多對D2D pairs該如何做到有效共用Channel的問題。首先考慮在理想狀況下通訊,系統可取得所有D2D信號與干擾的通道狀況。在此我們提出一套最佳排程公式,但我們發現在這公式中會遇到NP Hard的問題,於是無法找尋出一組最佳解。故在本文中提出了演化演算法的機制來找尋出一組次佳的解,透過此種機制使eNB能有效的對D2D Pairs進行排程。在理想狀況下,系統需做大量資料量測及收集大量資料,於是本論文提出了Location Based的量測方法,透過D2D之間的距離,來估計通道衰減可以有效的減少所需資料量測及資料交換。在本論文中有計算理想狀況下與Location Based的量測方法,從模擬結果中我們發現在Location Based與理想的情況下Outage Probability與Channel Usage 也有相近的效率。 Device to device make D2D pairs communicate without eNB in wireless communication system which is based on traditional cellular. It decrease burden of eNBs by decreasing jumping gain between D2D Paris and eNBs. And it effectively uses resource management to distribute resource which every D2D Pairs uses. According to the above methods, we can increase efficiency of wireless resource.In this paper, we discuss how to make many D2D pairs share channel effectively in LTE system. At first, we consider that system communicate in ideal situation. And system can get every D2D signal and interference situation. We submit an optimal schedule formula. But, we find it will encounter NP Hard problem. So, we can’t find an optimal solution. In this paper, we submit a mechanism about evolution algorithm to find a sub-optimal solution. By this mechanism, eNB can schedule D2D Paris effectively.In ideal situation, the system should measure and collect a large quantity of data. We can estimate channel fading by distance between D2D to effectively decrease data measurement and data swap. In this paper, we submit a computational method in ideal situation and a measure method of location based. According to the simulation result, we find that system in location based and in ideal situation, outage probability and channel usage have similar efficiency.