Top Banner
1 Spatio-temporal Allocation Takeshi Shirabe Technical University of Vienna
16
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: 1 Spatio-temporal Allocation Takeshi Shirabe Technical University of Vienna.

1

Spatio-temporal Allocation

Takeshi Shirabe

Technical University of Vienna

Page 2: 1 Spatio-temporal Allocation Takeshi Shirabe Technical University of Vienna.

2

• Spatial Allocation

• Mathematical Programming

• Spatio-temporal Allocation

Outline

Page 3: 1 Spatio-temporal Allocation Takeshi Shirabe Technical University of Vienna.

3

Spatial Allocation

Allocation of discrete spatial units (SUs)to larger groups or regionsaccording to criteria concerning such properties as size, shape,

and spatial relation.

Page 4: 1 Spatio-temporal Allocation Takeshi Shirabe Technical University of Vienna.

4

Spatial Allocation

Carve out regions!

Spatial Relation

Size & Shape

Size & Shape

Page 5: 1 Spatio-temporal Allocation Takeshi Shirabe Technical University of Vienna.

5

Spatial Allocation

Partition an area!

Page 6: 1 Spatio-temporal Allocation Takeshi Shirabe Technical University of Vienna.

6

Mathematical Programming

A methodology to formulate and solvedecision (programming or planning) problems

in algebraic terms.

Tool

Page 7: 1 Spatio-temporal Allocation Takeshi Shirabe Technical University of Vienna.

7

LP

If a problem is expressed as a set of linear equations,

we have a linear programming model.

For example…

Page 8: 1 Spatio-temporal Allocation Takeshi Shirabe Technical University of Vienna.

8

Problem: How to make a cup of the best cappuccino

by mixing coffee, milk, and foam?

LP

Page 9: 1 Spatio-temporal Allocation Takeshi Shirabe Technical University of Vienna.

9

Formulation:

252

0 , ,

50

3

21

321

321

x

xx

xxx

xxx 332211 xcxcxc

Constraints Objective function

Minimize or Maximize

ly.respective foam, andmilk cofee, of ][ amounts thebe and ,Let 321 mlxxx

LP

Page 10: 1 Spatio-temporal Allocation Takeshi Shirabe Technical University of Vienna.

10

0

x

bAx

cx

Solution:

LP

Feasible Region

polyhedron

Page 11: 1 Spatio-temporal Allocation Takeshi Shirabe Technical University of Vienna.

11

If a LP model requires one or more variables to be integer,

it is a (mixed) integer programming model.

IP

If you are lucky… In almost all cases…

Page 12: 1 Spatio-temporal Allocation Takeshi Shirabe Technical University of Vienna.

12

IP is a tool for spatial allocationsince SUs are finest or indivisible.

A decision is binary (0-1):whether or not each SU is allocated to a region.

Some work has been done by myself. Now…

IP for Spatial Allocation

Page 13: 1 Spatio-temporal Allocation Takeshi Shirabe Technical University of Vienna.

13

Spatio-temporal Allocation

It is concerned withwhere to do something andwhen to do itwhile spatial variation changes over time.

My interest.

Research has been going on elsewhere.

Page 14: 1 Spatio-temporal Allocation Takeshi Shirabe Technical University of Vienna.

14

• Static data and static decisions

(spatial allocation)• Dynamic data and static decisions

(e.g. facility siting)• Static data and dynamic decisions

(e.g. mining)• Dynamic data and dynamic decisions

(e.g. harvesting, routing)

Spatio-temporal Allocation

Page 15: 1 Spatio-temporal Allocation Takeshi Shirabe Technical University of Vienna.

15

Spatio-temporal Allocation

seems to be generalized to…

Allocation of discrete spatio-temporal units (?)to larger groups or regions (?)according to criteria concerning such properties as what????

Page 16: 1 Spatio-temporal Allocation Takeshi Shirabe Technical University of Vienna.

16

• Is ST allocation special or is it just a trivial extension of spatial allocation (e.g. ST allocation = 3D spatial allocation)?

• What kinds of properties are additionally involved?(e.g. contiguity in a (spatio-) temporal sense)

• Can they be decomposed into a set of primitive properties?

• Can such primitives be modeled in terms amenable to algorithmic solution (like IP)?

• Can they be recomposed into complex ST properties?

• Can GIS or DSS (decision support systems) benefit from this scheme?

Research Questions