Scribd
Upload
8 views5 pages

Advanced Incomplete Block Designs Class Notes

The document discusses various types of Balanced Incomplete Block Designs (BIBD), including Symmetric BIBD, Resolvable BIBD, and Affine Resolvable BIBD, along with their definitions, parameters, and important properties. It also covers concepts like Intra-Block Analysis, Complementary BIBD, Residual BIBD, Derived BIBD, and Dual BIBD, providing mathematical relations and examples for each. Additionally, it includes a summary of relationships between different designs and theoretical results such as Fisher's Inequality.

Uploaded by

sreedevipn
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
8 views5 pages

Advanced Incomplete Block Designs Class Notes

The document discusses various types of Balanced Incomplete Block Designs (BIBD), including Symmetric BIBD, Resolvable BIBD, and Affine Resolvable BIBD, along with their definitions, parameters, and important properties. It also covers concepts like Intra-Block Analysis, Complementary BIBD, Residual BIBD, Derived BIBD, and Dual BIBD, providing mathematical relations and examples for each. Additionally, it includes a summary of relationships between different designs and theoretical results such as Fisher's Inequality.

Uploaded by

sreedevipn
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

(Symmetric BIBD, Resolvable BIBD, Affine Resolvable BIBD, Intra-Block Analysis, Complementary BIB

Residual BIBD, Dual BIBD, Derived BIBD)

1. Symmetric BIBD
Definition
A BIBD is said to be Symmetric if:

v = b and r = k

Thus parameters become:

(v, v, k, k, λ)

Parameter Relations
From general BIBD relations:

vr = bk → vk = vk (automatically satisfied)

r(k − 1) = λ(v − 1)

Since r = k:

k(k − 1) = λ(v − 1)

Hence:

λ = k(k − 1)/(v − 1)

Important Property
In symmetric BIBD:

Any two blocks intersect in exactly λ treatments.

Proof Sketch:

Using duality and NN' = N'N structure.

1
2. Resolvable BIBD
Definition
A BIBD is Resolvable if the set of blocks can be partitioned into r groups (called replicates) such that:

• Each replicate contains v/k blocks • Each treatment appears exactly once in each replicate

Thus:

b = r(v/k)

Resolvable designs exist only if:

k divides v

Example
If v = 9, k = 3, r = 4

Then each replicate contains 9/3 = 3 blocks.

3. Affine Resolvable BIBD


An Affine Resolvable BIBD is a special resolvable design where:

• Any two blocks from the same replicate are disjoint • Any two blocks from different replica
constant μ treatments

Parameter relation:

μ = k²/v

Affine designs arise from finite affine geometries.

2
4. Intra-Block Analysis
Statistical Model
Y_ij = μ + τ_i + β_j + ε_ij

Constraints:

Σ τ_i = 0 Σ β_j = 0

Normal Equations
Treatment normal equations:

rτ_i − λ Σ τ_j = Q_i

Matrix form:

Cτ = Q

Where:

C = rI − λJ

Estimator of τ

τ̂ = C^− Q

Variance of difference:

Var(τ̂_i − τ̂_j) = 2σ²(r − λ)/rλ

Intra-block analysis adjusts treatment effects eliminating block effects.

5. Complementary BIBD
Given BIBD(v, b, r, k, λ)

The complementary design is formed by replacing each block with its complement.

New parameters:

3
v' = v b' = b k' = v − k r' = b − r λ' = b − 2r + λ

Proof: Count number of blocks where a pair occurs in complement.

6. Residual BIBD
Let original design be symmetric BIBD.

Choose one block and remove its treatments.

The design formed by remaining treatments and blocks is called Residual Design.

Parameters become:

v* = v − k b* =b− 1 r* = r − λ k* = k λ* =λ

7. Derived BIBD
Select one treatment and consider only blocks containing that treatment.

Remove that treatment from those blocks.

The resulting design is called Derived Design.

Parameters become:

v* =v− 1 b* = r r* = λ k* =k− 1 λ* =λ− 1

8. Dual BIBD
Interchange treatments and blocks.

If original design has parameters:

(v, b, r, k, λ)

Dual design has parameters:

(b, v, k, r, λ)

4
Dual of symmetric BIBD is again symmetric.

9. Relationship Summary Table


Original: (v, b, r, k, λ) Complementary: (v, b, b−r, v−k, b−2r+λ) Derived: (v−1, r, λ, k−1, λ−1
b−1, r−λ, k, λ) Dual: (b, v, k, r, λ)

10. Important Theoretical Results


1. Fisher’s Inequality: b ≥ v

2. For symmetric BIBD: b = v

3. Eigenvalues of Information Matrix: r − λ (multiplicity v − 1)

Examination Type Questions


1. Prove Fisher’s inequality.
2. Derive parameters of complementary BIBD.
3. Obtain parameters of derived design from symmetric BIBD.
4. Show dual of symmetric BIBD is symmetric.
5. Derive intra-block estimator of treatment effect.

You might also like