【正文】 ffect all the entries corresponding to previous nodes along the path. In order to facility smart updating, an ant must push the information about visited nodes into its stack: node identification, a binary mask that determines the status of free wavelengths on the links it traversed (this mask has W bits corresponding to the number of wavelengths). This stack also serves for loop detection and backtracking, to ensure that ants will not move forever on the network。 the best result of this approach is obtained through global update when the weight of attraction of ant for a path increases with the number of traversed ants. The result can be pared to the conventional Nagatsu heuristic [13], but it requires a much longer putational time. However, this approach cannot be applied directly to the dynamic RWA problem.Garlick et al. [14] proposed an ACObased algorithm to solve the dynamic RWA problem. When a new connection request arrives, a number of ants are launched from the source to the destination. Ants evaluate a path based on its length and the mean available wavelengths along the path. Global pheromone updating is performed when an ant reaches its destination. The pheromone updating is on a perdemand basis: the network pheromone matrix is reset once a connection is established. The final best path for a connection request is made when all ants plete their exploitation tasks. The authors showed that this algorithm has better performance than an exhaustive search over all available wavelengths for the shortest path [15]. As a new set of ants must be launched for each new connection request, the setup delay will be very high due to the waiting for ants in large networks. In fact, this approach does not show the collective behavior of ants that e from different requests, which is an important aspect of antbased systems。In this paper, we investigate a new antbased agent algorithm for the dynamic RWA problem in WDM networks under the constraint of wavelength continuity. Our study aims to reduce both blocking probability and path setup time by using a suitable amount of ants, which continuously perform path searching tasks before the connection request’s arrival so that the route selection and wavelength assignment of a request are performed by simply looking up the routing tables. To achieve that goal, we develop a new routing table structure, a scheme for ant population control and a mechanism for pheromone updating, for our new algorithm。In our work, we focus on the dynamic RWA problem with wavelength continuity constraint. In the literature, the dynamic RWA problem is usually divided into two subproblems that can be solved separately: routing and wavelength assignment. Routing schemes can be classified into fixed routing, fixedalternate routing and adaptive routing. In the fixed routing scheme, one route is dedicated for a sourcedestinationpair. Whenever a request occurs between this sourcedestination pair, this route is attempted for wavelength assignment. The fixed routing method is simple but usually causes a high blocking probability. The fixedalternate routing method has better performance with multiple paths dedicated for a node pair. In the adaptive routing scheme, the route is puted at the time the connection request arrives, based on the current network state, thus it yields the best performance. However, adaptive routing requires high putational plexity. A more detailed survey of routing and wavelength assignment can be found in [2]。對(duì)于最近一次訪問的相鄰i1節(jié)點(diǎn)，相應(yīng)的空閑波長(zhǎng)概率增加了，然而波長(zhǎng)對(duì)應(yīng)的概率繁忙程度降低了。當(dāng)螞蟻到達(dá)節(jié)點(diǎn)i，它將對(duì)應(yīng)節(jié)點(diǎn)s更新條目。當(dāng)一只螞蟻訪問一個(gè)節(jié)點(diǎn),它以其旅行過程中收集的信息來更新路由表的元素。當(dāng)一個(gè)連接請(qǐng)求到達(dá)時(shí)，路徑將決定基于最高的選擇概率相鄰節(jié)點(diǎn)的條目。當(dāng)一只螞蟻到達(dá)目的地節(jié)點(diǎn)或當(dāng)它不能選擇一個(gè)空閑的波長(zhǎng)選擇的路徑為其下一步行動(dòng)時(shí)將被剔除。 Ant launched Update pheromone Ant killed Fig. 2. Ant’s moving and updating tasks一只螞蟻從源移動(dòng)到目的地,在一Fig. 2. Ant’s moving and updating tasks個(gè)選定的波長(zhǎng)上一個(gè)節(jié)點(diǎn)到一個(gè)節(jié)點(diǎn)運(yùn)動(dòng)。這里ρ和T是設(shè)計(jì)參數(shù)。在這種情況下，因?yàn)镻1 P2，波長(zhǎng)2是優(yōu)于波長(zhǎng)1。 圖1所示的是當(dāng)W=1的一個(gè)新的路由表的新的例子。為了支持波長(zhǎng)分配,我們引入了選擇概率的每個(gè)波長(zhǎng)到路由表。每個(gè)條目對(duì)應(yīng)到目的節(jié)點(diǎn),每一列對(duì)應(yīng)一個(gè)相鄰節(jié)點(diǎn)。為了支持螞蟻路由選擇,每個(gè)網(wǎng)絡(luò)節(jié)點(diǎn)有一個(gè)路由表和N1條目。3. 基于蟻群的RWA算法 一個(gè)光學(xué)波分復(fù)用(WDM)網(wǎng)絡(luò)可以表示為由N個(gè)節(jié)點(diǎn)和E 鏈接的圖。作為一個(gè)新組螞蟻必須為新的連接請(qǐng)求啟動(dòng)，設(shè)置延時(shí)會(huì)由于大型網(wǎng)絡(luò)等待螞蟻而變得非常高。為一個(gè)連接請(qǐng)求的最后最好路徑的產(chǎn)生是當(dāng)所有的螞蟻完成他們的探索任務(wù)。當(dāng)一只螞蟻到達(dá)目的地，全球信息素更新被執(zhí)行。當(dāng)一個(gè)新的連接請(qǐng)求到來時(shí)，大量的螞蟻從源出發(fā)到目的地。然后，這個(gè)方法不能直接應(yīng)用于動(dòng)態(tài)RWA問題。該方法最好的結(jié)果是吸引螞蟻的路徑數(shù)量隨著穿越的螞蟻數(shù)量越來越多而獲得全球更新。每只螞蟻都留有一個(gè)用于路線回溯和循環(huán)回避的之前訪問節(jié)點(diǎn)的“禁忌”列表。一個(gè)螞蟻的路由選擇是基于每個(gè)連接的吸引力。目標(biāo)在于使一個(gè)給定網(wǎng)絡(luò)拓?fù)浜土髁烤仃嚨牟ㄩL(zhǎng)要求數(shù)量盡量減少。而以上的研究主要集中在電子通信網(wǎng)絡(luò)中的路由問題，我們?cè)诒疚闹械呐d趣是波長(zhǎng)連續(xù)性約束下的波分復(fù)用光網(wǎng)絡(luò)的動(dòng)態(tài)RWA問題。兩種基本算法是由Schoonderwoerd et al.[4]