Input data sources for

Uster, H. and Dalal, J., “Strategic Emergency Preparedness Network Design Integrating Supply and Demand Sides in a Multi-Objective Approach,” IISE Transactions, Vol. 49/4, pp. 395-413, 2017. (Featured in the ISE Magazine in March 2017.)

A. Computational study data

Note: The link to the problem data for “Network for Emergency Evacuation and Distribution of Supplies” (NEEDS) instances is given below. Please right click on the link and click “Save Target As” to save it on your machine.

• NEEDS Data Instances - NEEDS-DATASET.zip

All instance files are in “*.txt” format.

Format:

1. There are 8 different classes namely C1 to C8. Each class contains 5 network instances, and for each of these, there are 2 data files containing randomly generated model parameters. Therefore, each class contains 10 random test instances.

2. The network instances to represent the geographical region consisting of the evacuation zone, the potential shelter-opening zone, and the potential DC-opening zone, are named as “c1_a1.txt” to “c8_a5.txt”.

3. The data files associated with a network c(i)_a(j ).txt are named as c(i)_a(j )_n(k).txt, where i ∈ {1, ..., 8}, j ∈ {1, ..., 5}, and k ∈ {1, 2}. Therefore, we have data files “c1_a1_n1.txt” to “c8_a5_n2.txt”.

4. The Table 1 below gives the number of evacuation sources, shelters, and distribution centers (DC) in each class.

Table 1: Problem Classes for Computational Study

Class	Source (\|I\|)	Shelter (\|J\|)	DC (\|K\|)	Total (\|N\|)
C1	2500	200	50	2750
C2	2500	200	60	2760
C3	2500	250	50	2800
C4	2500	250	60	2810
C5	3200	200	50	3450
C6	3200	200	60	3460
C7	3200	250	50	3500
C8	3200	250	60	3510

5. The legend used in the input network files c1_a1.txt to c8_a5.txt are self-explanatory. Those are mentioned in Table 2.

Table 2: Legends used in Network instance files

Symbol	Details
[HabitatNodelist : loc_id,x,y]	List of (x,y) coordinates of all habitats (evacuation source) nodes.
[ShelterNodelist : loc_id,x,y]	List of (x,y) coordinates of all candidate Shelter nodes.
[DCNodelist : loc_id,x,y]	List of (x,y) coordinates of all candidate DC (supply) nodes.

6. The legend used in the data file are explained in Table 3.

Table 3: Legends used in data instance files

Symbol	Details
[Habitat Population : i, pop_i]	Population at evacuation source i
[Shelter Population : j, pop_j]	Population at candidate Shelter j
[Product : id, demand per head, transport cost factor]	For each relief supply p, demand per evacuee, and unit DC-to-shelter transportation cost
[MaterialFlow_VarCost : j,VC_ j]	Variable cost at shelter j to handle relief supplies from DC
[DC Inv Holding Capacity : k,p,S {k,p}]	Capacity of DC k to hold inventory of relief item p
[DC Opening Cost : k, f_k]	Fixed cost of opening DC at k
q_{j,l}	Number of evacuees who can get accommodation within level l of shelter j
f_{j,l}	Fixed cost of opening a shelter at j with capacity level l
VC_{j,l}	Variable cost to serve evacuees in level l of shelter j

Figure 1: Setting for test instance generation

B. Case study data

For the case study discussed in Section 6 of the paper, we use publicly available GIS data as shown in Table 4.

Table 4: Data used in case study with sources

Data	Source
Census 2010 population data for each evacuation origin (block-group level) and each candidate shelter (county level)	www. socialexplorer.com
Census 2010 TIGER/Line GIS shapefiles having “block-group” and “county” as geographic units to construct the evacuation, shelter and DC zones as given in Table 4.	www.census.gov

We use the following counties given in Table 5, to construct the evacuation, shelter, and DC zones respectively.

Table 5: Counties considered in three zones in our GIS case study

Evacuation Zone

Shelter Zone

DC Zone

Selected block groups of the following 5 counties, closely resembling the evacuation map by Texas DPS: Brazoria, Chambers, Galveston, Harris, Matagorda.

Anderson, Angelina, Bastrop,

Bell, Bexar*, Blanco, Bosque,

Brazos*, Burleson, Burnet,

Caldwell, Cherokee, Comal,

Coryell, Dallas*, Ellis,

Falls, Freestone, Gillespie,

Gregg, Guadalupe, Hamilton,

Harrison, Hays, Henderson, Hill,

Houston, Johnson, Kaufman,

Kendall, Lampasas, Lee, Leon,

Limestone, Llano, Madison,

Mason, McLennan*, Milam,

Mills, Nacogdoches*, Navarro,

Panola, Robertson, Rusk, San

Saba, Shelby, Smith*, Tarrant*,

Travis*, Trinity, Upshur,

Vanzandt, Walker*, Williamson,

Wood. (56 counties)

Archer, Baylor, Bowie, Brewster,

Callahan, Clay, Coke, Cooke,

Crane, Crockett, Edwards, Fannin, Fisher, Glasscock, Grayson,

Haskell, Howard, Irion, Jones,

Kinney, Knox, Lamar, Martin, Maverick, Midland, Mitchell,

Montague, Nolan, Pecos, Reagan, Real, Red river, Schleicher, Scurry, Shackelford, Sterling, Stonewall, Sutton, Taylor, Terrell, Throckmorton, Tomgreen, Upton, Uvalde, Val verde,

Wichita, Young, Zavala. (48

counties)

Source of evacuation: Centroid of Block Group.

Candidate Shelter:

Centroid of County.

Candidate DC:

Centroid of County.

*Counties considered as shelters in the “State Evacuation and Sheltering Plan” (Texas, 2006)

Case study : input data files and structure

· 1-Source-Location.xlsx

Source ID (i)

Zone

Col 1: source node ID (i)

Col 2: longitude (x) of source node i

Col 3: latitude (y) of source node i

Col 4: which zone source node “i” belongs to: Coast/A/B/C.

· 2-Source-Population.xlsx

Source ID (i)

Population

Col 1: source node ID (i)

Col 2: source node i’s population (census 2010)

· 3-Shelter-Location.xlsx

Shelter ID (j)

Col 1: shelter node ID (j)

Col 2: longitude (x) of shelter node j

Col 3: latitude (y) of shelter node j

· 4-Shelter-Misc.xlsx

Shelter ID (j)

Population (pop_j)

VC^M_{j}

Col 1: shelter node ID (j)

Col 2: shelter node j’s population (census 2010)

Col 3: variable cost for relief handling at shelter j

· 5-Shelter-Capacity-Cost.xlsx

Capacity_Level (l)

Shelter ID (j)

Capacity (q_{jl})

Fixed Cost (f_{jl})

Variable Cost (vc_{jl})

Col 1: shelter capacity level (l): small (1) / medium (2) / large (3)

Col 2: shelter node ID (j)

Col 3: capacity of the shelter at level “l”.

Col 4: fixed cost of opening shelter “j” at capacity level “l”

Col 5: variable cost of operating shelter “j” at capacity level “l”

· 6-DC-Location.xlsx

DC ID

Col 1: DC node ID (k)

Col 2: longitude (x) of DC node k

Col 3: latitude (y) of DC node k

· 7-DC-Fixed-Cost.xlsx

DC ID (k)

fixed cost (f_k)

Col 1: DC node ID (k)

Col 2: fixed cost of opening DC at node “k”

· 8-DC-Capacity.xlsx

DC ID (k)

Relief Item ID (p)

Capacity, S_{kp}

Col 1: DC node ID (k)

Col 2: relief item ID (p)

Col 3: capacity of DC “k” to store relief item “p”

· 9-Miscellaneous.xlsx

(p)

Per Head Demand

unit transportation cost

Col 1: relief item ID (p)

Col 2: per head demand for item “p”

Col 3: Unit transportation cost for item “p” ($/unit-mile)

Unit Transportation Cost for evacuees, α

Col 1: unit transportation cost of evacuees ($/person-mile)