非經常性災難疏散之動態廣播策略

Dynamic Broadcast Strategies for Non-recurrent Disaster Evacuation

黃晟哲(Cheng-Che Huang)

指導教授 ： 坂井勝哉

國立臺灣大學/工學院/土木工程學研究所/碩士(2020年)

https://doi.org/10.6342/NTU202002871

摘要

資訊傳播對於非經常性災難疏散是不可輕視的一環，因而許多的研究致力於探索資訊傳播如何影響著非經常性災難之交通狀況。雖然這些研究聚焦於不同之資訊提供策略，例如：資訊傳播速率、資訊傳播手段等，然而充足之編列預算總是被考慮。如此一來，資訊傳播的速度設置總是為固定的並且資訊在整個疏散期間可以被充分地傳播。然而，其他研究指出由於事前預算編列之限制，相關應變機構應存在著預算限制。受到這些研究的啟發，一動態資訊提供策略應被考慮、建構、實施以期待有效地使用有限的預算。
在這篇研究中，我們探討在非經常性災難之狀況下廣播如何影響災難疏散之效益。此外，因為事前預算是被提前編列給相關應變機構，故廣播存在著有限之預算。而在資訊網路中，影響疏散之旅行資訊是以以下二種方式傳播：一種是駕駛與駕駛間之旅行資訊傳播；另一種則是廣播對於駕駛之旅行資訊傳播。藉此，路網中的駕駛則會根據他們所接受到之旅行資訊以做出各自的路徑選擇。而根據有限的廣播預算，這篇研究提供了兩種不同的動態廣播策略，分別稱為“快慢策略”和“慢快策略”。此外，為了幫助決策者規劃動態廣播策略，我們還提出了三個假說。這些假說考量了路線選擇機會分佈、旅行時間分佈和疏散需求，並通過數值模擬進行驗證。之後，我們使用資訊傳播模型套入Sioux-falls路網以進行交通模擬，除了分析這兩種廣播策略之差異，還分析了不同廣播頻率設置的影響。更進一步地，我們分析了兩種廣播策略中較佳之策略並給予廣播頻率不同之設置。結果表明，於Sioux-falls路網中，實施“快慢”廣播策略具有較佳的疏散效益。然而，如果採用相當極端的“快慢”廣播策略，疏散效益也可能會變差。

關鍵字

疏散 ； 交通模擬 ； 社會網路 ； 資訊傳播 ； 動態資訊提供 ； 廣播策略

並列摘要

A lot of research has studied how traffic state during non-recurrent disaster changes with information propagation. This is because information propagation usually plays a pivotal role in society to influence disaster evacuation. Although these studies focus on different information provision strategies concerning information propagation speeds and information propagation tools, a sufficient budget for disasters are only considered. The speed of information propagation is set to fixed, and information is sufficiently propagated for the whole evacuation period. However, other research indicates a budget limit exists for emergency agencies because of the ex-ante budget. Inspired by these studies, a dynamic information provision strategy should be implemented to use finite budgets in an efficient way.

In this study, we focus on how broadcast influences evacuation performance for non-recurrent disaster. Also, we considered the budget limit for the broadcast. The budget limit results from the cost for disasters which are usually budgeted by emergency agencies in advance. Travel information is transmitted on the information network in two ways. One is drivers-to-drivers information, and the other is broadcast information. According to the travel information they have received, drivers have to use them and make route choices. Two different dynamic broadcast strategies which are named as “fast-slow” and “slow-fast” are provided according to finite broadcast budget. To help decision-makers to plan dynamic broadcast strategies, three hypotheses considering route choice opportunities distribution, travel time distribution, and evacuation demand are proposed and verified by numerical simulation. And we conduct traffic simulation with an information propagation model to analyze the effects of the two broadcast strategies and different setup of broadcast frequency in the Sioux-falls network. Further, we analyze different setups of broadcast frequency for the better strategy of the two broadcast strategies. The results indicate that the “fast-slow” broadcast strategy should be implemented to have better evacuation performance. However, when extreme “fast-slow” broadcast strategies are adopted, evacuation performances may get worse.

並列關鍵字

Evacuation ； Traffic simulation ； Social network ； Information propagation ； Dynamic information provision ； Broadcast strategy

參考文獻

Acar, A., Muraki, Y. (2011). Twitter for crisis communication: lessons learned from Japan's tsunami disaster. International Journal of Web Based Communities, 7(3), 392-402.

Baker, E. J. (1991). Hurricane evacuation behavior. International journal of mass emergencies and disasters, 9(2), 287-310.

Barrett, B., Ran, B., Pillai, R. (2000). Developing a dynamic traffic management modeling framework for hurricane evacuation. Transportation Research Record, 1733(1), 115-121.

Bish, D. R., Sherali, H. D., Hobeika, A. G. (2014). Optimal evacuation planning using staging and routing. Journal of the Operational Research Society, 65(1), 124-140.

Charnkol, T., Tanaboriboon, Y. (2006). Tsunami evacuation behavior analysis: One step of transportation disaster response. IATSS research, 30(2), 83-96.