Using GIS, Genetic Algorithms, and Visualization in Highway Development
Authors:
Jha, McCall, & Scholfeld
Instructor: Prof Crouch
Presenter: Mike Jones
Problem:
How can we best leverage Information Technology to improve the planning process for highway development in order to achieve the optimal balance of cost, schedule, and performance?
Motivation:
In Maryland alone, cost overruns in highway projects cost $297.6 M in 1999.
This problem is not unique to Maryland. Do you remember the Coliseum Central Highway Improvement Project?
Approach:
A blend of GIS, GA, and CV.GIS: Geographic Information Systems
Data WarehouseGA: Genetic Algorithms
OptimizationCV: Computer Visualization
Feedback and based in intangiblesSecure stakeholder support
Geographic Information System:
Mines data from various sources to provide a concise, easily understood representation.Aerial photographs.MDPropertyView
Raster Property maps and attribute information.Maryland State Highway Administration
Soil layers, floodplains, wetlands
Geographic Information Systems:
Mines data from various sources to provide a concise, easily understood representation.
Compute costs based on:Right of way.Environmental concerns.
Geographic Information Systems:
Mines data from various sources to provide a concise, easily understood representation.
Compute costs based on:Right of way.Environmental concerns.
By assigning high cost to environmentally sensitive areas
Geographic Information Systems:
Mines data from various sources to provide a concise, easily understood representation.
Computed costs consider:Agency costs: Right of way, environmental,
pavement, construction, maintenance, & earthwork.
User costs: accidents, travel time, & vehicle operation
Accounting note:
One time cost per unit length: construction, maintenance & pavement.
5-year period cost: accidentsNPV in base year: fuel, travel time
NO CONSISTENT METHOD FOR COMPARISON!!
Optimization
Classic technique – derivatives 2D example:Assume an initial solution.Take derivative. If derivative is positive, decrease estimate
and repeat If derivative is negative, increase estimate
and repeat If derivative is zero, optimal solution found!
Genetic Algorithms:
In the class of global search heuristics called evolutionary algorithms.
Use multiple initial guesses, called the initial population Evaluate the fitness of each individual in the population Repeat
Select best-ranking individuals to reproduce Breed new generation through crossover and mutation
(genetic operations) and give birth to offspring Evaluate the individual fitnesses of the offspring Replace worst ranked part of population with offspring
Until <terminating condition>
Recursive solution:
The cost calculated by the GIS is the fitness criteria used in the GA. Ideally, the GA will converge to the optimal solution.
Stakeholder commitment
Once the optimal solution is reached, visualization will be used to educate and earn the commitment of key stakeholders:ManagementCitizensLegislators (funding)