## Pavement Management Decision Analysis Using Belief Functions in Valuation-Based Systems

### BibTeX

@MISC{Attoh-okine_pavementmanagement,

author = {N. Attoh-okine},

title = {Pavement Management Decision Analysis Using Belief Functions in Valuation-Based Systems},

year = {}

}

### OpenURL

### Abstract

Valuation-Based Systems (VBS) for belief-functions theory is applied to Pavement Management Systems (PMS) decision-making. The VBS provides a general framework for representing knowledge and drawing inferences under uncertainty. A VBS network is constructed and potentials are introduced in the form of belief-function (or basic probability assignment) in PMS decision making environment. Valuation network is another method of representing and solving Bayesian decision problems. It is based on the framework of VBS. Valuation network depict decision variables, random variables, utility functions, and information constraints. The solution method for valuation network is called fusion algorithm, and the Dempster's rule of combination can be successfilly applied in this framework. It will be shown that this approach can capture the quantitative, qualitative and incomplete information in PMS decision making. 1.

### Citations

55 |
A valuation-based language for expert systems
- Shenoy
- 1991
(Show Context)
Citation Context ...el, associated with the bpa p is defined by: Another way of expressing the information contained in bpa function p is in terms of the plausibility function PL: PL(a) = C(p(b) 1 bnaZ0) for each aeZwh. =-=(4)-=- Intuitively, the plausibility of a is the degree to which a is plausible in the light of the evidence. A zero plausibility for a hypothesis means that we are sure that it is false, but a zero degree ... |

10 |
Bellman: Dynamic Programming
- E
- 1957
(Show Context)
Citation Context ...ies is obtained by : (QD {r1,.*.r,,pI,...pn} (+I (1) The optimal strategy U* tlhat gives us the maximum expected value of A is determined as follows: 40 (nQDp)'D(d) = (:@ {PI ,... r,,p1,... pn> ) (+) =-=(2)-=- where A, p, and D refers to the equivalent canonical decision problem 4. The solution method for VBS is based on fusion algorithm (3). The basic iidea is to successively delete all variables from the... |

2 |
P.: Valuation Networks, Decision Trees, and Influence Diagrams: A Comparison
- Shenoy
- 1993
(Show Context)
Citation Context ...ned as follows: 40 (nQDp)'D(d) = (:@ {PI ,... r,,p1,... pn> ) (+) (2) where A, p, and D refers to the equivalent canonical decision problem 4. The solution method for VBS is based on fusion algorithm =-=(3)-=-. The basic iidea is to successively delete all variables from the VBS. The sequence in which the variables are deleted must respect the precedence constraint. The following theorem describes the fusi... |

1 |
Attoh-Okine, "Potential Use of Valaution-Based systems and Networks in Pavement Management Systems
- unknown authors
- 1994
(Show Context)
Citation Context ...ets of W,,,, and cE{(atkUh)nbtgUh))}. If both ui and uj are belief functions, then their combination is defined by Dempster's rule of combination: q@uj = KC(pi(Xlh,a) pj(X'g,b) 1 at@uh)nbt@uhl = cl ; =-=(5)-=- where K = 1 - C(pi(a)pj(b) 1 atkuh)nbt(guh) = 0 1 If neither of ui and uj are belief functions, then ui8uj is obtained by: C(pi(Xlh,a) + pj(XAg,b) I a'(puhlnb'(guh) = CF) Otherwise, vi @ uj is obtain... |

1 |
A fusion Algorithm for Solving Bayesian Decision Problems
- Shenioy
- 1991
(Show Context)
Citation Context ...iables from the VBS. The sequence in which the variables are deleted must respect the precedence constraint. The following theorem describes the fusion algorithm for making inferences in VBS. Theorem =-=(6)-=- Suppose A = {X,,L {WJ,,,, (~1..."1, {p l...pm],+], is a well-defined decision problem. Suppose X,X 2...X, is a sequence of variables in X=XD U X, such that wit!h respect to the partial orders3. Belie... |

1 |
A Decision calaculus for Belief Functions in Valaution Based Systems
- Xu
(Show Context)
Citation Context ... neither of ui and uj are belief functions, then ui8uj is obtained by: C(pi(Xlh,a) + pj(XAg,b) I a'(puhlnb'(guh) = CF) Otherwise, vi @ uj is obtained by: 1 {pi(Xhh,a) pj(Xlg,b) I atkuh) nbt(guh) = c} =-=(7)-=- The combination has the following properties: (a) Commutativity (b) Associativity (c) If both ui and uj are belief functions, then ui@uj is also belief function (d) If both are utility valuations, th... |