D Nagesh Kumar, IISc Optimization Methods: M6L1
1
Dynamic Programming Applications
Design of
Continuous Beam
Objectives
¾
To discuss the design of continuous beams
¾
To formulate the optimization problem as a dynamic
programming model
D Nagesh Kumar, IISc Optimization Methods: M6L1
3
Design of Continuous Beam
¾ Consider a continuous beam having n spans with a set of loadings W1, W2,…, Wn
¾ Beam rests on n+1 rigid supports
¾ Locations of the supports are assumed to be known
W1 0
L1
W2 1
L2
2
Wi i-1
Li
Wi+1 i
Li+1
i+1
Wn n-1
Ln
n
Design of Continuous Beam …contd.
¾ Objective function: To minimize the sum of the cost of construction of all spans
¾ Assumption: Simple plastic theory of beams is applicable
¾ Let the reactant support moments be represented as m1, m2,
…, mn
¾ Complete bending moment distribution can be determined once these support moments are known
¾ Plastic limit moment for each span and also the cross section of the span can be designed using these support moments
D Nagesh Kumar, IISc Optimization Methods: M6L1
5
Design of Continuous Beam …contd.
¾ Bending moment at the center of the ith span is -WiLi/4
¾ Thus, the largest bending moment in the ith span can be computed as
¾ Limit moment for the ith span, m_limi should be greater than or equal to Mi for a beam of uniform cross section
¾ Thus, the cross section of the beam should be selected such that it has the required limit moment
n i
L for W m
m m m
Mi i i i i i i 1,2,...
4 , 2
,
max 1 1 =
⎭⎬
⎫
⎩⎨
⎧ + −
= − −
Design of Continuous Beam …contd.
¾ The cost of the beam depends on the cross section
¾ And cross section in turn depends on the limit moment
¾ Thus, cost of the beam can be expressed as a function of the limit moments
¾ Let X represents the vector of limit moments
⎪
⎪⎪
⎬
⎫
⎪
⎪⎪
⎨
⎧
= m m
X _lim lim _
2 1
Μ
D Nagesh Kumar, IISc Optimization Methods: M6L1
7
Design of Continuous Beam …contd.
¾ The sum of the cost of construction of all spans of the beam is
¾ Then, the optimization problem is to find X to Minimize
satisfying the constraints .
¾ This problem has a serial structure and can be solved using dynamic programming
∑
= ni
i X
C
1
) (
∑
= ni
i X
C
1
) (
n i
for M
m_ limi ≥ i =1,2,...,