This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: 16.410/413 Principles of Autonomy and Decision Making Lecture 16: Mathematical Programming I Emilio Frazzoli Aeronautics and Astronautics Massachusetts Institute of Technology November 8, 2010 E. Frazzoli (MIT) L07: Mathematical Programming I November 8, 2010 1 / 23 Assignments Readings Lecture notes [IOR] Chapters 2, 3, 9.13. [PA] Chapter 6.12 E. Frazzoli (MIT) L07: Mathematical Programming I November 8, 2010 2 / 23 Shortest Path Problems on Graphs Input: V , E , w , s , G : V : set of vertices (finite, or in some cases countably infinite). E ⊆ V × V : set of edges. w : E → R + , e → w ( e ): a function that associates to each edge a strictly positive weight (cost, length, time, fuel, prob. of detection). S , G ⊆ V : respectively, start and end sets. Either S or G , or both, contain only one element. For a pointtopoint problem, both S and G contain only one element. Output: T , W T is a weighted tree (graph with no cycles) containing one minimumweight path for each pair of startgoal vertices ( s , g ) ∈ S × G . W : S × G → R + is a function that returns, for each pair of startgoal vertices ( s , g ) ∈ S × G , the weight W ( s , g ) of the minimumweight path from s to g . The weight of a path is the sum of the weights of its edges. E. Frazzoli (MIT) L07: Mathematical Programming I November 8, 2010 3 / 23 Example: pointtopoint shortest path Find the minimumweight path from s to g in the graph below: Solution: a simple path P = s , a , d , g ( P = s , b , d , g would be acceptable, too), and its weight W ( s , g ) = 8. E. Frazzoli (MIT) L07: Mathematical Programming I November 8, 2010 4 / 23 Another look at shortest path problems Cost formulation The cost of a path P is the sum of the cost of the edges on the path. Can we express this as a simple mathematical formula? Label all the edges in the graph with consecutive integers, e.g., E = { e 1 , e 2 , . . . , e n E } . Define w i = w ( e i ), for all i ∈ 1 , . . . , n E . Associate with each edge a variable x i , such that: x i = 1 if e i ∈ P , otherwise. Then, the cost of a path can be written as: Cost ( P ) = n E i =1 w i x i . Notice that the cost is a linear function of the unknowns { x i } E. Frazzoli (MIT) L07: Mathematical Programming I November 8, 2010 5 / 23 Another look at shortest path problems (2) Constraints formulation Clearly, if we just wanted to minimize the cost, we would choose x i = 0, for all i = 1 , . . . , n E : this would not be a path connecting the start and goal vertices (in fact, it is the empty path). Add these constraints: There must be an edge in P that goes out of the start vertex. There must be an edge in P that goes into the goal vertex. Every (non start/goal) node with an incoming edge must have an outgoing edge A neater formulation is obtained by adding a “virtual” edge e from the goal to the start vertex: x = 1, i.e., the virtual edge is always chosen....
View
Full
Document
This note was uploaded on 12/26/2011 for the course SCIENCE 16.410 taught by Professor Prof.brianwilliams during the Fall '10 term at MIT.
 Fall '10
 Prof.BrianWilliams
 Mass

Click to edit the document details