<div dir="ltr"><div dir="ltr"><div dir="ltr"><div dir="ltr"><div dir="ltr"><div dir="ltr"><font color="#000000" size="4" style="" face="times new roman, serif">Sayin Liste Uyeleri, </font><div><div><font color="#000000" size="4" style="" face="times new roman, serif"><br></font></div><div><font face="times new roman, serif" size="4"><font color="#000000" style="">Gebze Teknik Üniversitesi (GTU) Matematik Bölümü Genel Seminerleri kapsamında, 22 Kasım Cuma günü saat 14:00'te Dr. </font><span style="color:rgb(0,0,0)">Anıl Yazıcı</span> (<span lang="EN-US" style="">Stony Brook Üniversitesi)</span><font color="#000000" style=""> bir seminer verecektir. Seminerin detayları aşağıda olup tüm ilgilenenler davetlidir.</font></font></div><div><font color="#000000" size="4" style="" face="times new roman, serif"><br></font></div><div><font color="#000000" size="4" style="" face="times new roman, serif">Saygılarımızla. </font></div><div><font color="#000000" size="4" style="" face="times new roman, serif"><br></font></div><div><font color="#000000" size="4" style="" face="times new roman, serif">Dear all,</font></div><div><font color="#000000" size="4" style="" face="times new roman, serif"><br></font></div><div><font face="times new roman, serif" size="4"><font color="#000000">There will be a seminar in Gebze Technical University (GTU) on 22nd of November by Dr. </font><span style="color:rgb(0,0,0)">Anıl Yazıcı</span> (<span lang="EN-US" style="">Stony Brook

University)</span><font color="#000000">.</font></font></div><div><font color="#000000" size="4" style="" face="times new roman, serif"><br></font></div><div><font color="#000000" size="4" style="" face="times new roman, serif">Time and place: At 14:00 in Department of Mathematics, Lecture Amphi 2</font></div></div><div><font color="#000000" size="4" style="" face="times new roman, serif"><br></font></div><div><font face="times new roman, serif" size="4"><font color="#000000">Title: </font><span lang="EN-US" style="">Link Criticality Index (LCI) for Analysis of

Large Transportation Networks</span></font></div><div><font color="#000000" size="4" style="" face="times new roman, serif"><br></font></div><div><font face="times new roman, serif" size="4"><font color="#000000">Abstract: </font> Designing and

maintaining resilient and robust transportation systems require identification

of links that are critical for the functionality of the network. Network

scientists have suggested various topological measures to identify critical

network components in multiple domains such<span style="letter-spacing:-2.05pt">

as energy, water supply, telecommunication, social, and transportation.

However, topological criticality measures usually do not conform to traffic

flow dynamics (e.g., non-linear link performance<span style="letter-spacing:-0.5pt"> </span>functions)<span style="letter-spacing:-0.6pt"> </span>and<span style="letter-spacing:-0.55pt"> </span>they<span style="letter-spacing:-0.8pt"> </span>do<span style="letter-spacing:-0.55pt"> </span>not<span style="letter-spacing:-0.55pt"> </span>necessarily<span style="letter-spacing:-0.65pt"> </span>produce<span style="letter-spacing:-0.6pt">

</span>realistic<span style="letter-spacing:-0.6pt"> </span>results<span style="letter-spacing:-0.45pt"> </span>for<span style="letter-spacing:-0.65pt"> </span>road<span style="letter-spacing:-0.55pt"> </span>transportation networks. <span style="letter-spacing:-0.15pt">In </span>order to overcome this issue,

researchers incorporate traffic characteristics (travel time, link flow, etc.)

into criticality metrics, mostly through traffic assignment. The studies that

employ traffic assignment generally utilize a network performance measure

(e.g., total system travel time) for the complete network, remove each link one

by one, re-run traffic assignment<span style="letter-spacing:-1.55pt"> </span>and

calculate<span style="letter-spacing:-0.7pt"> </span>the<span style="letter-spacing:-0.6pt"> </span>selected<span style="letter-spacing:-0.65pt"> </span>network<span style="letter-spacing:-0.65pt"> </span>performance<span style="letter-spacing:-0.7pt"> </span>measure<span style="letter-spacing:-0.75pt"> </span>for<span style="letter-spacing:-0.65pt"> </span>each<span style="letter-spacing:-0.55pt"> </span>link<span style="letter-spacing:-0.65pt"> </span>removal,<span style="letter-spacing:-0.55pt">

</span>and<span style="letter-spacing:-0.65pt"> </span>identify<span style="letter-spacing:-0.8pt"> </span>each<span style="letter-spacing:-0.65pt"> </span>link’s

criticality based on the change in the network performance measure. Such full

network scan approach creates computational burden, especially for large

networks. Link removals can also cause<span style="letter-spacing:-0.4pt"> </span>network<span style="letter-spacing:-0.4pt"> </span>disconnectivity<span style="letter-spacing:-0.65pt"> </span>which<span style="letter-spacing:-0.35pt"> </span>makes<span style="letter-spacing:-0.35pt"> </span>it<span style="letter-spacing:-0.3pt"> </span>problematic<span style="letter-spacing:-0.4pt"> </span>to<span style="letter-spacing:-0.35pt"> </span>run<span style="letter-spacing:-0.4pt"> </span>traffic<span style="letter-spacing:-0.4pt">

</span>assignment.<span style="letter-spacing:-0.35pt"> </span>This<span style="letter-spacing:-0.35pt"> </span>talk<span style="letter-spacing:-0.35pt"> </span>will

present the recently developed Link Criticality Index (LCI) which does not

require link removals and can provide transportation link criticality rankings

for large networks with a single User Equilibrium<span style="letter-spacing:-0.6pt"> </span>(UE)<span style="letter-spacing:-0.65pt"> </span>traffic<span style="letter-spacing:-0.5pt"> </span>assignment.<span style="letter-spacing:-0.55pt"> </span>The<span style="letter-spacing:-0.55pt"> </span>LCI<span style="letter-spacing:-0.75pt"> </span>scores<span style="letter-spacing:-0.6pt">

</span>are<span style="letter-spacing:-0.55pt"> </span>calculated<span style="letter-spacing:-0.6pt"> </span>based<span style="letter-spacing:-0.6pt"> </span>on<span style="letter-spacing:-0.5pt"> </span>the<span style="letter-spacing:-0.65pt"> </span>link<span style="letter-spacing:-0.6pt"> </span>flow<span style="letter-spacing:-0.65pt"> </span>changes

and<span style="letter-spacing:-0.45pt"> </span>link<span style="letter-spacing:-0.45pt"> </span>performance<span style="letter-spacing:-0.35pt"> </span>functions<span style="letter-spacing:-0.4pt"> </span>during<span style="letter-spacing:-0.45pt">

</span>UE<span style="letter-spacing:-0.35pt"> </span>iterations.<span style="letter-spacing:-0.3pt"> </span>In<span style="letter-spacing:-0.45pt"> </span>order<span style="letter-spacing:-0.35pt"> </span>to<span style="letter-spacing:-0.4pt"> </span>take<span style="letter-spacing:-0.5pt"> </span>the<span style="letter-spacing:-0.35pt"> </span>alternative<span style="letter-spacing:-0.35pt"> </span>OD<span style="letter-spacing:-0.2pt"> </span>paths<span style="letter-spacing:-0.4pt"> </span>into account, the links scores are

weighted based on the travel time differences between identified OD paths.<span style="letter-spacing:-0.3pt"> </span>Also,<span style="letter-spacing:-0.3pt"> </span>to<span style="letter-spacing:-0.3pt"> </span>reflect<span style="letter-spacing:-0.3pt">

</span>the<span style="letter-spacing:-0.2pt"> </span>higher<span style="letter-spacing:-0.25pt"> </span>importance<span style="letter-spacing:-0.35pt"> </span>of<span style="letter-spacing:-0.25pt"> </span>links<span style="letter-spacing:-0.3pt"> </span>which<span style="letter-spacing:-0.3pt"> </span>serve<span style="letter-spacing:-0.3pt"> </span>multiple<span style="letter-spacing:-0.35pt">

</span>ODs<span style="letter-spacing:-0.3pt"> </span>(or<span style="letter-spacing:-0.3pt"> </span>higher<span style="letter-spacing:-0.35pt"> </span>demand ODs),

demand weights are used. The LCI was tested on well-known large transportation

test networks<span style="letter-spacing:-0.65pt"> </span>and<span style="letter-spacing:-0.55pt"> </span>compared<span style="letter-spacing:-0.55pt">

</span>with<span style="letter-spacing:-0.6pt"> </span>other<span style="letter-spacing:-0.75pt"> </span>criticality<span style="letter-spacing:-0.9pt"> </span>measures.<span style="letter-spacing:-0.65pt"> </span>The<span style="letter-spacing:-0.6pt"> </span>findings<span style="letter-spacing:-0.55pt">

</span>showed<span style="letter-spacing:-0.65pt"> </span>that<span style="letter-spacing:-0.55pt"> </span>LCI<span style="letter-spacing:-0.8pt"> </span>successfully

provides<span style="letter-spacing:-0.4pt"> </span>balanced<span style="letter-spacing:-0.4pt"> </span>link<span style="letter-spacing:-0.45pt"> </span>criticality<span style="letter-spacing:-0.6pt"> </span>rankings<span style="letter-spacing:-0.4pt">

with<span style="letter-spacing:-0.4pt"> respect<span style="letter-spacing:-0.4pt"> to<span style="letter-spacing:-0.4pt"> network<span style="letter-spacing:-0.45pt"> connectivity<span style="letter-spacing:-0.65pt"> and<span style="letter-spacing:-0.45pt"> redundancy<span style="letter-spacing:-0.65pt"> as well as the network flow conditions.

Due to elimination of link removals, the LCI runtime is considerably<span style="letter-spacing:-2.1pt"> </span>shorter than comparable traffic

assignment based approaches, i.e., in the order of the number of links in the

network. The inclusion of network topology and demand characteristics, and<span style="letter-spacing:-0.65pt"> </span>shorter<span style="letter-spacing:-0.7pt">

</span>runtime<span style="letter-spacing:-0.7pt"> </span>make<span style="letter-spacing:-0.6pt"> </span>the<span style="letter-spacing:-0.6pt"> </span>LCI<span style="letter-spacing:-0.7pt"> </span>an<span style="letter-spacing:-0.65pt"> </span>appealing<span style="letter-spacing:-0.65pt"> </span>approach<span style="letter-spacing:-0.65pt">

</span>for<span style="letter-spacing:-0.7pt"> </span>transportation<span style="letter-spacing:-0.65pt"> </span>link<span style="letter-spacing:-0.65pt"> </span>criticality<span style="letter-spacing:-0.8pt"> </span>analysis.</font></div></div></div></div></div></div></div>