The optimization of emergency resource-mobilization based on harmony search algorithm

Transcription

1 Avalable ole Joural of Chemcal ad Pharmaceutcal Research, 04, 6(7): Research Artcle ISS : CODE(USA) : JCPRC5 The optmzato of emergecy resource-moblzato based o harmoy search algorthm Le Xog Da School of Mathematcs ad Computer Scece, Shaa Uversty of Techology, Hazhog Shaa, Cha ABSTRACT Ths paper establshes a optmzato model of emergecy resource-moblzato based o the prcple of the shortest tme ad mmum cost, whch s a obectve fucto. I order to calculate ths optmzato problem, harmoy search algorthmc s used to solve the problem. Epermetal results show our proposed method s effectve ad effcet for solvg the mult-obectve optmzato problem. Key words: harmoy search; mult-obectve optmzato; emergecy resources moblzato ITRODUCTIO Emergecy resource-moblzato composes of may kds of emergecy resources ad emergecy uts. Ad the dstrbuto of varous emergecy uts s depedet ad rregular. Emergecy resources of all emergecy uts are lmted. The dstace betwee emergecy uts ad pot s dfferet. Therefore, speed ad cost of resource-moblzato affect the schemes drectly. At preset, research o resource schedulg s mostly lmted to the costrats of schedulg model ad route choce, ad for large scale problems, there s less solvg algorthm of may cases. I case of emergecy pots less, ehaustve method ca geerally obta the optmal soluto of the problem. However, whe more emergecy resources partcpate, ehaustve method solves ths kd of problems hopeless. So we employ the heurstc optmzato algorthm to solve ths problem. Harmoy search algorthm [8] s a ew heurstc search algorthm ad has successfully bee appled to the structure optmzato desg, traffc path plag ad the slope stablty aalyss, etc. Research shows that the algorthm dscrete optmzato problems s better tha the at coloy algorthm, geetc algorthm performace. The epermetal results show that harmoy search algorthm has better performace solvg such problems. ESTABLISH THE OBJECTIVE FUCTIO Let S,S,,ad S are alteratve resource supply pot, s the th( =,,,m)effectve rescue resources of S. (=,,, ). Assumg that, a moblzato of resources, the demad value of th s a,ad = a,from the avalable resources to provde emergecy resources selected p (p <= ) resource pots volved rasg, Ad meet the "rase the shortest tme" ad "to rase the cost of at least" two goals. If the tme from rase resources pot S to place A s t (t > 0),ad t < t < < t. Accordg to "rase the shortest tme", the ca establsh the obectve fucto s: 483

2 Le Xog Da J. Chem. Pharm. Res., 04, 6(7): f = m T ( ) = ma( t ) =,,..., If we cosder the cost of rasg resources, amely, rase the eed aother key obectve codtos: rescue resources to rase the mmum cost. From the resources pot set S to rase emergecy resource to reach the ut cost of the rally pot A requred for c, the lowest cost kd of rescue resources spet at the: C = = c y "The obectve fucto s to meet the mmum fee" for the: f = MI( C ) = MI c y = I the solvg process, accordg to "rase the mmum cost" of ths goal, ca be coverted to "rase the mmum" problem, order to meet the varous emergecy resource moblzato of resources, makg out at least some. HARMO SEARCH ALGORITHM Harmoy search (Harmoy Search, HS [8-]) algorthm s a ew tellget algorthm 00 of a South Korea scholar Geem Z proposed by W et al. The three ma parameters of the algorthm, amely the harmoy memory cosderg probablty HMCR (Harmoy memory cosderg rate), PAR (Ptch adustg rate trmmg probablty), the ptch adustg badwdth BW (bad wdth).the basc steps of harmoy search algorthm s as follows: Step: Settg of parameters The umber of varables : ; Rage of each varable; 3 Mamum teratos: Tma; 4 Sze of harmoy memory : HMS; 5 Harmoy memory cosderato rate: HMCR; 6 Ptch adustg rate: PAR; 7 Badwdth: bw. Step: Italzg the harmoy memory (HM) radomly, the HM s as follows, H M f ( ) L f ( ) f ( ) L f ( ) = = M M M M L M M H M S H M S H M S H M S H M S H M S f ( ) L f ( ) Step 3: mprovsg the ew Harmoy. The geerato of ew harmoy, = (,, L, ),the ew harmoy (=,,,) ca create through the followg three kds of mechasm: The learg ad memory; the toe tug;3the radom selecto of toe. That s: HMS (,, ) L If rad<hmcr = =,,, X, otherwse; Where X deotes the value of the th varable, rad represets a radom uform dstrbuto o [0,], For from the harmoy database ad toe tug to the probablty of PAR,.e.: 484

3 Le Xog Da J. Chem. Pharm. Res., 04, 6(7): Where rad represets a radom uform dstrbuto o [0, ], PAR as the ptch adust probablty, BW s toe tug badwdth. Step 4: Update memory Evaluato of the ew geerato of harmoy, f ew harmoy s better tha the worst the harmoy memory harmoy, harmoy wll be updated to the ew harmoy memory. For eample, to fd the optmal path,.e.: If + rad * bw = ( k + m ), m {,}, otherw se; f f f ( ) < ( worst ) = ma ( ), =,, L, HMS If rad<par (Cotuous type) If rad<par (Dscrete type) the worst = Step 5: f the termato codto s reached, output the optmal soluto. Otherwse, go to Step 3. EXPERIMETAL SECTIO EXPERIMETAL SIMULATIO Problem descrpto Assume that d (=), d (=9), d 3 (=36) are three essetal resources the rescue process, there are 8 resources rescue pot (S, S S 8 ). The per cost that th rescue resources s trasport to the accdet pot A s C. All cost of resource rescue pot s show table ; Reserves of th resources rescue pot for the, s show Table. The shortest tme t of arrvg accdet locato from resources rescue pot s show table 3. Table the cost from resources rescue pot to the locato of the accdet =,,, C S S S S 3 S 4 S 5 S 6 S 7 S 8 C C C Table the resource reserve scale of resources rescue pot S S S S 3 S 4 S 5 S 6 S 7 S Table 3 the shortest tme that arrve accdet locato from resources rescue pot t S S S S 3 S 4 S 5 S 6 S 7 S 8 t Solvg the problem by HS algorthm Accordg to table, table, ad table 3, mult target applcatos to set up to rase model, we employ HS algorthm to solve the emergecy resource-moblzato. The HS algorthm s tally appled to solve cotuous optmzato problems, solvg combatoral optmzato problems, ad s used to geerate a ew post. There may be llegal solutos, resultg HS algorthm caot be used, therefore, the applcato of HS algorthm combato optmzato problems, must realze ts dscretzato. I the soluto process, the key parameters: the probablty HMCR=0.8, ptch adustg probablty PAR=0., umber of mamum teratos MAXITER=000, toe tug badwdth of BW=0., ad the memory sze of HMS=5, solvg process as show fgure : 485

4 Le Xog Da J. Chem. Pharm. Res., 04, 6(7): Italzato Geerate a radom umber Rad Rad<HMCR Select a scheme S the tal lbrary Radomly geerate a scheme S Wth PAR tug S S s better tha the Worst soluto? Update Termate? Output results Fg. flow char of harmoy search algorthm The results s obtaed by usg the MATLAB 7 programmg algorthm to solve the problems metoed above, the epermetal results are as follows: Table 4 resource moblzato scheme C S S S 3 S 4 S 5 S 6 S 7 S 8 /C /C 4/5 / 5/8 4/3 0/0 0/0 6/5 0/0 /C 6/5 0/0 /5 6/0 0/0 0/0 4/0 /8 3/C 3 7/4 5/8 7/6 0/0 7/ 0/0 0/0 0/0 Accordg to the followg table rasg scheme, three kds of resources ca be obtaed by rasg scheme, t as follows, The scheme for resource d : {(S, 4), (S, ), (S3, 5), (S4, 4), (S7, 6)}; The scheme for resource d : {(S, 6), (S3, ), (S4, 6), (S7, 4), (S8, )}; The scheme for resource d 3 : {(S, 7), (S, 5), (S3, 7), (S4, 0), (S5, 7)}. The shortest respose tme t of resource d s MAX (t) = (ut tme); the lowest cost of d s 486

5 Le Xog Da J. Chem. Pharm. Res., 04, 6(7): =4 (ut cost); The shortest tme of resource d s 8 (ut tme), the lowest cost of d s 8 (ut cost). The shortest tme of resource d 3 s 6(ut tme), the lowest cost of d 3 s 87 (ut cost). Therefore, the rescue process, the earlest start tme of the emergecy rescue should be the mamum respose tme of three resources, t s 6 (ut tme). COCLUSIO Ths paper adopts the harmoy search algorthm to solve the problem of emergecy resources. The epermetal results show that the harmoy search algorthm has advatages of fast covergece speed, whch s smple ad tutve, less parameters, easy to mplemet. The method s effectve ad feasble to solve emergecy resources moblzato. Ackowledgemet Ths paper s supported by the fud proect of Shaa Provce Educato Offce (03JK6); Shaa Uversty of Techology foudato proect (slgky3-7) REFERECES [] Al H Sefchago, Trasportato Research Part a Polcy ad Practce, 996, 30():3-50. [] George F. Lst, IEEE Trasactos O Egeerg Maagemet, 998, 45(): 4-5. [3] Fedrch F, Gehbauer F ad Rckers U, Safety Scece, ():4-57. [4] Mea L;Hatao Che, Operatos Research ad Maagemet Scece, 0, 0 (3): [5] Guoqag We;L Jg, Statstcs & Decso 00, 0:0-. [6] Je Zhag;Zhyog Wag, Computer Egeerg ad Applcatos, 0, 47 (3): 0-3. [7] Jgu Jag;Jgguo Lag, Statstcs & Decso,009, : [8] Geem ZW;Km JH;Logaatha GV, Smulato,00, 76(): [9] Shouheg Tuo;Faga Deg, Computer scece, 0, 09: [0] Tuo Shouheg;Logqua og ad Tao Zhou. Mathematcal Problems Egeerg 03, [] Shouheg TUO;Logqua, Joural of Computatoal Iformato Systems, 0, 8(0): [] Shouheg TUO;Logqua OG & Fag a DEG,04, 487