Hazardous Waste Logistics: Strategic and Tactical Decisions - PowerPoint PPT Presentation

1 / 47
About This Presentation
Title:

Hazardous Waste Logistics: Strategic and Tactical Decisions

Description:

Logistics management of a RSW collection and disposal system with the following stages: ... Risk to zone Zk by travel on a = Risk acceptability and equity. Each ... – PowerPoint PPT presentation

Number of Views:144
Avg rating:3.0/5.0
Slides: 48
Provided by: MSi62
Category:

less

Transcript and Presenter's Notes

Title: Hazardous Waste Logistics: Strategic and Tactical Decisions


1
Hazardous Waste Logistics Strategic and
Tactical Decisions
Department of Transport Engineering Pontificia
Universidad Católica de Chile
Rodrigo A. Garrido
2
Starting with an easier problem...Design of an
Urban Waste Transport System
  • Santiago is the capital of Chile with 6 million
    inhabitants.
  • 1980s Santiago generated 0.63 kg/Inhab-day of
    residential solid waste today it reaches 1.4
    kg/Inhab-day.
  • Currently 95,000 tons of RSW are generated in
    Santiago, which are disposed into 2 Sanitary
    Landfills

3
Problem Definition
  • Logistics management of a RSW collection and
    disposal system with the following stages
  • Collection picking up waste from generation
    points
  • Hauling transporting it from generation points
    or Transfer Centers to landfill
  • Final disposal dropping off the load into the
    landfill

4
Simplification Zoning the City

5
Simplification Assuming fixed routing cost

6
The problem to solve is to...
  • Locate unknown number of Transfer Centers and
    Landfills legal regulation
  • Distribute the waste to TC and from these to the
    Landfills
  • Choose type of trucks for both journeys
  • These decisions are simultaneously made to
    minimize the total system cost

7
Proposed Solution Mixed integer programming
model
  • Objective Function includes installation and
    operating costs for both stages
  • Subjected to capacity constraints, available
    technology, demand satisfaction and non
    negativity.

8
Application to Santiago
  • Planning horizon year 2020
  • Annual growth rate of 3
  • Centroids reflect demographic density
  • 14 possible landfill sites and 5 for TC
  • 240 Integer variables, 19 Binary variables
  • 240 Continuous variables, 589 Constraints

9
Heuristic Solution Network partition and
penalties
  • But...If the waste was dangerous?

10
Why do we need to move Hazmat?
  • Hazmat include explosives, flammable, oxidant,
    poisonous, infectious, radioactive, or corrosive
    substances among others
  • Industrial production processes use hazmat as
    manufacturing components and often generate
    hazmat by-products.
  • Hazmat must be safely transported from their
    origin to special facilities to be used in a
    production process, modified to decrease their
    degree of danger, or properly disposed.

11
Generation and Distribution of Hazardous Waste in
an Urban Context
  • Objectives
  • To estimate the generation of hazwaste in a
    metropolitan area
  • To develop a model for routing hazwaste in the
    urban transportation network
  • To locate treatment plants

12
How to proceed...?

13
The Generation Model
  • Industrial classification into 28 Categories
  • Standard hazmat 14 categories
  • - Corrosive, Toxic, Reactive, Flammable.
  • Current generation ???
  • Forecasting aggregated ARIMA model
  • Desegregation at county level, Land Use
  • Hazwaste to be transported per county

14
Disaggregation at county level

15
How much of that is transported?

16
What is the cost of an incident with hazmat?
  • Potential to provoke fatalities or injuries to
    the inhabitants in the vicinity of the event
  • Impact area defined by a dispersion radius ?
    that depends substance properties and topography

17
An accident in an arc and its consequence
Probability of an accident in route R
18
Now, lets route!!!

Sherali et al (1997)
19
(No Transcript)
20
Problems with that formulation
  • If we had several hazmat types?
  • How much of m, h is too much? not yet...
  • If we find an optimal route,
  • what happens when used daily?
  • What is the meaning of a fair solution?

21
Routing multiple hazmat frequently transported
  • Each with different degree of danger

22
But... did you forget about the people?
  • Consider an arc a and a hazmat with radius l

23
How to incorporate equity on routing?
  • Control the difference in risk for any two zones
  • Lets call it the Gopalans constraint
  • Gopalans constraint can be included in the
    routing model we have so far

24
(No Transcript)
25
How can we improve Gopalans constraint?
  • It assumes a fixed acceptable risk for each zone
    How high that risk could be? not yet...
  • It assumes that the hazmat movement will produce
    the same maximum risk on each zone
  • Do all the zones have the same baseline risk?

26
The Relative Degree of Risk
  • Degree of Risk for given hazmat r

27
A practical value for the risk function
  • Assuming a constant risk function w/r to space in
    Gopalans expression

people

28
Risk acceptability and equity
  • Each zone has a basline risk Bk

We can modify P1 with this new equity constraint
29

30
Multiobjective Routing of Multiple Hazmat

31
How else could we improve the system?
  • So far we have found a set of non dominated paths
    to route hazmat
  • Each shipment travels from the generation point
    to a fixed and known treatment plant
  • How many plants should we have?
  • Where should they be located?

32
OK, but Not in my backyard!
  • Locating hazmat treatment plants require at
    least
  • Low accident probabilities at each site
  • Low acc prob during transportation to each site
  • Low consequences if an accident occurs
  • Low total logistics cost
  • High social acceptability
  • High equity
  • Clearly a multiobjective location problem

33
Some Definitions
  • In a directed graph, let O be the set of hazmat
    generators and D the set of potential plants

34
Objectives of our Model
  • Minimum assignment cost
  • Minimum setup cost
  • Min catastrophic consequences
  • Expected number of affected trucks
  • Cost of catastrophic accident in site j Caj
  • Expected consequence in j

35
The Model
36
Multicriteria Analysis Multiobjective
Programming
  • Proceed in two steps
  • Search for a set of feasible solutions Pareto
    optimal
  • Multicriteria function to find Best Solution

37
First Step Finding the Pareto Optimal Set
  • For a given vector (a1, a2, a3) solve

38
Second Step The best compromise-solution
  • For a given metric p solve
  • L1 gives the best compromise solution for
    linear deviation
  • L2 gives the Euclidean distance
  • L8 gives the most equitable

39
Application Hazwaste in Santiago
Chile(ongoing!!!)
  • Santiago has 39 boroughs.
  • The strategic transportation network has 6,000
    arcs and 2,000 nodes.
  • Generation in 24 centroids, with 6 potential
    treatment plants
  • Four types of hazwaste

40
(No Transcript)
41
Santiagos Hazwaste CaseEstimation of Accident
Probabilities
  • Summary of estimated conditional probabilities at
    network level 

42
Santiagos Hazwaste CaseEstimation of the
Consequences
Impact area for each category of hazmat
43
Santiagos Hazwaste CaseEstimation of the
Consequences
Summary of registered accidents by type of
hazmat between years 1993-1999
44
The probability of an accident
  • Identify and evaluate scenarios for the
    occurrence of an incident involving hazmat
  • Scenarios can be described as a sequence of
    events. The events can be represented in a tree
    structure, showing a sequential progression of
    branched options event tree

45
Outcome of the event treeProb of a
catastrophic accident in arc a
46
Final Remarks
  • A lot of unresolved issues in literature
  • Risk measurement
  • Equity
  • Routing and location jointly?
  • Bi-level optimization problem?
  • Interactions between hazmat?
  • Non-constant costs?
  • Road pricing for hazmat, what to charge?

47
Any questions?
whos that guy?
Write a Comment
User Comments (0)
About PowerShow.com