P MODELING MULTI-HIERARCHY EMERGENCY

leading charac of the polygon generator is n spatial dissect density in the setting points. There are W. and Gu traditional one p3……pn} in Ti={x|dx, pi distance. The w providing that plane, an λ λ 、λ 、 d p, pi λj ⁄ , i the weight of t 2 Multi-ga matching Accordin the road netw diagram and r stack. We can the overlap lay the polygon constructed by the zone whe gathering poin selecting some Accordin road evacuati thickness and established by multi-settling Tab.1 is the s relationship o points, we ca has four adjac three adjacent adjacent point cter is that the d n composing by nearest. We u tion concerning gathering poin e many ways to Ch. L., 2000; e is provided n the plane, s dx, pj p weighted one is t n discrete po nd n λ … … , Vn j , divided t the point. athering points g ng to the coordi ork structure, w oad network cl analyses the sp yer. As shown proximity rela y every point. In ere there is a nt and the admi e setting points Figure 1. O ng to the road ion capacity a d the weighte y the spatial loc points, we can Table 1. Spa space adjacency of the multi-ga an see from the ent points here points here: s here: T4, T distance betwee y some points se weighted V g the populati nts and the adm o generate Voro ; Zhen Ch. H that n discrete so the relevant pi, pi pj}. d s a little differe ints P={p1, p2 positive n Pi, λi ∩ the plane into s and multi-se inates of the sim we construct th lassified layer, a patial point mat in the followin ations in the n accordance w a point, the p it population in to every gather Overlaying laye network layer as a group of ed Voronoi d ation of multi-g obtain the follo ace adjacent tab y table about th athering point e table that th e: T10, T11, T1, T2, T 5, T6, T7; S en the internal in the plane an Voronoi diagram on and distrib mit population i onoi diagram Z H.et al. 2011 . e points P={p1 t Voronoi dia denotes Eucli ent from it. Like 2, p3……pn} in real num P| d p, pi λi ⁄ n parts, λi den ettling points s mulating points he weighted Vo and then make tching in the lig ng fig.1, we ca radioactive re with road capac population in e n every setting p ring point. rs which describe f lines in diff iagram whic gathering point owing table: ble he space simul s and multi-se he setting poin T12, T13; S T3; S3 has S4 has four adj point nd its m as bution in the Zhang The 1, p2, agram idean ewise, n the mbers i notes space s and ronoi them ght of n see egion ity in every point, es the ferent ch is ts and lation ttling nts S1 2 has four acent poin topo and the T base and acco we c S1 a T13 matc shou popu free betw the r T12 can poin the s

3.2 P

setti to se rally part cond obta over seve gath to M one beca to th over max et al 2000 gath eval optim selec poin secti secti unde we s nts here: T ological relation the weights ab Tab. 2 as below Table Tab.2 is ultim ed on the topolo the traffic ommodation o can see that, and the adopt ca which shou ching setting po uld distribute t ulation accept population acc ween the origi remaining popu , T11 has bee define the corr nt and can get t setting points. Path searching To transfer th ng points, we h elect the optima y points to settle in constructing dition. Accord ained based on rlay of road n eral setting poin hering point, an Multi-destinatio path as evacua ause one path o he road transpo rall evacuation ximum flow on l.2009; Ch J. Ch 0, selecting hering point t uation criteria mal K path. T ction method is nt to other setti ions. According ions, we dete erstanding path select the path T7, T8, T nships of th bout the ability w. e 2. Spatial poin mate space-point ogical relationsh c capacity of population o according to apacity of each ld consider f oint of S1 are: the population tance of poin eptance of T1 nal population ulation after the en completed . responding sett the number of g he disaster vict have several rou al path to make e safely and eff g emergency ev ding to the s n the weighted etwork layer, nts that is con nd that is the pa ons. Usually it’ ation route from ften causes the ort pressure, an time. We use t the premise of h. et al. 2007; A the optimum o single setti of road transp The basic thou s that there are s ing points, and g to the discre ermine the o h from gathering of least resistan T9, T10.Acco he various gat of the road traf nt matching tabl t matching pro hips of the two of road netw of each point . F the simulated setting points: firstly, we ca T11, T12, T13 to the best pl nt T12, T11 13 is the diffe n acceptance o e transmission According to ting points of the population ims from gath utes to choose f e the disaster vi ficiently is the m vacuation model spatial point m d Voronoi figu we get some sidered prefere ath searching o ’s impractical m a rally point short time cong nd directly caus the K optimal f global optimu Ahuja R K and m several path ing point acc portation ability ught of the op several roads fr d each road is etion of the imp overall impeda g points to setti nce as optimal ording to the thering points ffic, we can get le gram which is kind of points work and the From the table population of T10, T11, T12, an define the 3, Because we lace, when the is full, the erence number of T13 and between S1 to the method we each gathering distribution of ering points to from. And how ictims from the most important l under disaster matching way ure and layers information of entially in each of single Origin if only choose to another site, gestion, adding se the delay of paths based on um Kou W. H. Magnanti T L, hs from single ording to the y, namely, the ptimal K path rom a gathering made by some pedance of the ance of each ing points, then path, which in e s t s , e e f , e e e e r d o e g f o w e t r y s f h n e , g f n . , e e e h g e e h n n XXII ISPRS Congress, 25 August – 01 September 2012, Melbourne, Australia 197 turn to the se again until K o 3.2.1 Multi-pa constraints: T section is base flow, road dam realize the Rea process of eva path method b of increasing t caused by sh population dis node, as below Figure 2. The secti of the impeda transportation the weight o population on Based on available to ev point. Three p 2, 4, 7, 9}; {1 nodes is alread relationship am people in the f Road sectio Road tran ability Distributed population Based on following form people of each Calculation o nodes: calculation law econd best sele optimal path. aths problem he assessment ed on the inform mage degree an al-time optimal acuation under based on the ma traffic flow and hort term perso stribution inform w. .Population di ions weight in f ance of those r capacity, the a of each section the section from n the above fig vacuate people paths, respective 1, 2, 5, 8, 9}; { dy available in mong road sec figure, the follo ons 1-2 nsportation a1 Table 3 n the ideas o mula to calcula h path: of the popula ; the po ; the po ; th w, the populatio ection, the opt m under the t of traffic ca mation of real-t d static road co l quality of sele r disaster, we a aximum flow t d affecting overa onnel congestio mation of each istribution base figure is assess roads which is a1, a2, a3......s n. indica m link node 1 t gure, starting th from the gathe ely, go through 1, 3, 6, 9}, roa data of road ne ctions, nodes an owing table is ob 1-3 2-4 a2 a3 . Information ta f maximum f ate the final num ation flow be pulation flow b opulation flow he population ; then on on each road timal path sele e maximum apacity of sel time dynamic tr onditions. In ord ected road durin adopt the K op o solve the pro all evacuation s on. We recode h section in the ed on max-flow ed by the recip s to be assessed eparately stand ates the distrib o node 2. here are three p ering point to se different nodes ad sections with twork. Accordi nd the high flo btained: 4 2-5 … a4 … … able flow, we adop mber of distrib etween 1-2 n between 1-3 n between 2-4 n flow between according to d section is obta ection flow ected raffic der to ng the ptimal oblem speed e the links procal d the ds for buted paths etting s: {1, h link ing to ow of … … … t the bution nodes: odes: odes: n 2-5 the ained. A popu certa Duri we w the flow reco mini 3.2.2 effic theo area infor algo desti well typic node by s foun optim trave searc data are p vario algo high road evac netw mod princ by t H={ and of ro as: w base thres ζ After the pop ulation needs to ain settlement p ing the assignm will record the node informati w on a path, orded in all th imum value as 2 Road search ciency when pa ory to calculate a 、 real time rmation road orithms could ination point w l if the question cal shortest pat es of the shorte step search algo nd from them to mal solution, b ersal, so ineffic ching algorithm a sets. To solve proposed Xian ous networks, t orithms are diffe hest efficiency t d network and cuation model work hierarchy del for path op ciples are show Fig All road netw the ability of {H1,H2,H3,H4} V means the se oad, W means th Vi v , w , w … … . ed on the abil shold , pulation on ea o be identified point on a cert ments process b flow of people ion. In order t you should co he nodes inclu the final popula hing in networ ath searching p multiple optim and includin network data. find the optim while those alg n is very compl th algorithm to est path. Becaus orithm for each o m shortest pat but because of cient. That is t m suites to sma this problem, s ng J. P. et al.20 the efficiencies ferent .And ther to all kinds of n d the detailed under the d searching is u ptimizing searc wn in the figure gure 3 .Road ne work data is div road evacuatio }. The graph of et of road node he set of weigh v , v , v … … .The threshold lity of road e value: ; ζ ach section is from an assem tain accessible based on the m e assigned to ro to determine th ompare the po uded in the p ation flow on th rk hierarchy: I processes based mal paths whic ng POIpoint . Traditional p mal path from gorithms could licated. Dijkstra o calculate a no se Dijkstra algo h vertex n rese th to work, it c f its calculatio to say, Dijkstr all data sets ra some optimal ro 11; Qiao L. et s of all optimal re is no algorith networks. Based requirements disaster condit used to establis hing. The who below. etwork layers vided into four on which is re f road network es, E means the ht of each sectio ; Ei e , e , value of each l evacuation. As ζ ζ , ζ , ζ , ; ζ s figured out, mbly point to a road transport. maximum flow, oad sections in he final person opulation flow ath, then take he entire path. It will has low d on max-flow h face to large of interesting path searching start point to dn’t figure out a algorithm is a ode to all other orithm is a step ervations so far an arrive at the on of the node a shortest path ather than large oute algorithms al. 2011 .As to route planning hm that has the d on the special of emergency tion, the road sh mathematics ole algorithmic specific layers epresented as : k G={V, E, W} set of sections on of road, such e … … ; Wi layer is defined ssume that the , ζ . , a . , n n w e w w e g o t a r p r e e h e s o g e l y d s c s : s h d e XXII ISPRS Congress, 25 August – 01 September 2012, Melbourne, Australia 198 , ; ζ ; . For each layer Hi∈ H, we compare Wiv, e with threshold ζi and make sure the current road data belongs to which layer. After all road data has compared with threshold, we will get four sets such as: Hi= {Vi, Ei, Wi}, i=1, 2, 3, 4. Searching the K optimal paths from gathering point to setting point based on following steps: Step1: Finding the nearest road to the current gathering point and compare the ability of road evacuation with threshold of each layer; Step2: Looking for the other collinear points with current gathering point and compare those points with threshold value to make sure they are located in which layer. Assume that the set of road nodes of collinear road section be represented as: N={n|ni ∈ Ls, i , … ….}. Step3: Taking one node from the set N to begin the next step of path searching. Step4: If the current road node is located in the highest layer then search the next section of road which is collinear with it. The end node of the current section of road as the current node, that is, to reuse the same method over and over again until the current node is collinear with settling point. Step5: If we can’t find the collinear node with current node in the current data layer, we need to find the mapping point of current node in the next layer and reuse Step 4 and Step 5 until to find all the successful way from all gathering points to all setting points. The concrete methods are shown in the figure below. Figure 4. Detailed network

3.3 Vehicle Assignment