UNIVERSITI PUTRA MALAYSIA

STATE LEVEL LOG PRODUCTION PLANNING USING LINEAR PROGRAMMING

ABDUL RAHMAN ABDUL RAHIM

FH 1989 1

STATE LEVEL UX; PROOOCI'ICN PLANNING USING LINEAR P.ROORAMMING

By

ABOOL RAHMAN ABOOL RAHIM

Thesis Submitted in Partial Fulfilment of the Requirements for the Degree of Master of Science

in the Faculty of Forestry Universit i Pertanian r.Blaysia

June 1989

In the name of Allah , the rrost Benevolent and Merciful.

All praise be for Allah for giving Ire the opportunity,

patience and guidance in canpleting this thesis. My deepest

gratittrle and love to my family, my wife, Soriati , and sons ,

Abdullah and Badruzzaman, who constantly pray for my success.

I w ish to express my appreciation and sincere gratitude to

my major supervisor , Dr. Yusuf Hadi and co-supervisor En. Awang

Noor Abd. Ghani , fran the Faculty of Forestry , Universiti

Pertanian Malaysia for their assistance and guidance

throughout the duration of this study.

t-ly deepest appreciation to the Director General , Forestry

Department of Peninsular Malaysia who made this study possible.

The financial support from the Public Services Department

(Jabatan Perkhidmatan Awam) Malaysia is gratefully

ackncmledged •

I am also grateful to the Dean, Faculty of Forestry,

Uni versi ti Pertanian t-1alaysia , for granting the use of

facilities for the study.

Last rut rot least , I VX)uld like to extend my appreciat ion

to all those who have contributed to the success of this study.

May Allah bless them all. Amin.

11

TABLE CF aNI'ENl'S

L Isr OF TABLES

LIsr OF FIGURES •

L Isr OF ABBREVIATIOOS •

ABSTRAcr

ABSTRAK •

I IN'IROOOcrIOO •

Problem statement

•

• •

Objective� of the Study )

Scope of the study •

•

•

•

Organization of the Report •

II REVIEW OF L ITERA'IURE • • •

• •

•

•

Methods of Log Production Planning .

Area Control

Volume Control

• • •

Ccrnbination of Area-Volume Control

Simulation

Linear Programning

• •

•

•

•

Other Forestry App lications of Linear Programming in Ma.laysia •

iii

Page

ii

vi

viii

ix

xi

xiv

1

1

5

5

6

8

8

8

9

10

11

12

15

Page

Desirable Features for Log Proouction Plarming �1odel • • • • • • • • • • • • • • • • • • • 16

Synthesis of Log Production Planning in r1:llaysia • • • • • • • • • • • • • • • • • • • • 16

III HETOOIX)ILGY . . . . . . . . . · . . . . . .

Data Collection and Generation • . . . . . . . .

r-bdel Developnent . . . . . . . . .

OVerall Structure • . . . . . . .

L inear Prograrrrning r.1odel Developnent

Standtable Projection r.1odel Development •

Application of the Proposed Nodel to Negeri Sembilan • • • • • • • • • · .

Detennination of Log Production • • · . . . . . .

Sensitivity Analysis . . . . . . . . . . . .

IV RESULT . . . . . . . . . . . . . . .

18

18

19

19

20

31

41

41

43

45

The Proposed Planning t10del 45

The Overall Structure of the Proposed P lanning r.1odel • • • • • • • • • • • • • • • 45

Linear Prograrrming r-bdel

Stand Projection Model

Operation of the Pro{:X)sed Model •

Yields from the Forests

Log Proouction • • • • • • . . . .

Sensitivity of Future Log Production for Negeri Sembilan • • • • • • • •

iv

. . .

· . . . . . .

46

50

55

56

61

63

v

Effect of Additional CUtting Cycle

Relative Importance of Forest Areas

. . .

DISaJSSION • • • • • • • • • • • . . . . . . . . . .

Special Features of the Prq::osed Planning Model • • • • • • • • • . . . . .

Ccxnparison of Forecasted U::lg Prcduction

Areas for Further Research • • • •

Implication of the Forecasted Log Production • • • • • • • • • • • •

. . . .

VI SUMMARY . . . . . . . . . . .

Log Prcduction Planning Model . . . . . . . . . . .

U:>g Production frc.:xn Negeri Sembilan . . . . .

Reccmnendations . . . . . . . . . . . . . . . .

REFERENCES . . . . . . . . . . . . . . . . . . . . . . .

AOOITIOOAL REFERENCES . . . . . . . . . . . . . . . . .

APPENDICES

A Linear Programming Problem Matrix for Sc:erta.r io 2 • • • • • • • • • • • • • • • •

B Standtable Projection

. .

Page

63

65

70

70

7 4

78

82

8 4

84

85

87

89

96

1 00

102

VITA • • • • • • • • .. • • • • • • • • • • • • • • • • • 127

v

Table

1

2

3

4

5

6

7

8

9

LIST OF TABLES

Extent of Forest Areas in Negri Semb ilan • • • • • • • • • • • • • • • • •

LOruS 1-2-3 Spreadsheet layout . . . . . . . . .

Fonnat of Standtable . . . . . . . . . . . . . .

Sunmary of the two Scenarios . . . . . . . . . .

'!he Problem Matrix for Scenario 1 • • • • • • • •

Conversion of Stock ing to Cumulat ive Stock ing in Init ial Stand • • • • • • . . . . . .

Distribut ion of Tree Stock ing in Nonna lized Stand • • • • • • • • . . . . . . . .

Standtable Project ion for Superior Forest Stand • • • • • • • • • • •

Range Name and Def init ion U sed in WillS 1-2-3 • • • • • • • • . . . . . . . . .

1 0 Estimates of furtali ty Rate I Basal Area

1 1

12

and Voltml9s • • • • • • • • • • • • • •

Future Log Prcduct ion Pattern for Negeri Semb ilan (1990-2060) • • • • • • • • • •

Area to be Harvested in the Future For Negeri Semb ilan (1990-2060) • • • • • •

vi

. . . . .

Page

19

32

36

42

47

50

52

53

56

57

61

64

Table

13

14

15

16

17

18

19

I..oY.er and Upper Lind t s and Incentive Values far Scenario 1 • • • • • • • •

I..oY.er and Upper Lind t s and Incentive Values far Scenario 2 • • • • • • • •

Variables with ' Penalty Cost ' in

. . . . . .

. . . . . .

Scenario 1 • • • • • • • • • • • . . . . . . . .

Variables with ' Penalty Cost ' in Scenario 2 • • • • • • • • • • • . . . . . . . .

Canparison of Forecasted Log Prcrluction (million cu . m per year )

Area Control U sing rata U sed in LPPM

Canpar ison of Existing and Future Stand in LPPH and Current Practice (cu. m per hal • • • • • • • • • •

vii

. . . . . . .

. . . . . . .

Page

66

67

68

69

75

77

79

Figure

1

2

3

4

5

6

LIST OF FIGURES

Structure of the Log Product ion Planning t-'k:xJ.el ( LPPM ) • • • • • • • • • • • • • • . . . .

The Ba sic Steps in Stand Proj ect ion • . . . . . .

The Trends of Yields for the Forest s

Basal Area for all Forest Types •

Forecasted Log Production Pattern for Negeri Sembilan • • • • • • • •

Canparison of Forecasted Log Volume Production • • • • • • • • • • • •

viii

. . . . . .

. . . . . . .

. . . . . . .

Page

2 1

34

58

60

62

76

LIST OF ABBRE.VIATICNS

Abbreviation

AIFM

cu. rn

an

dbh

FAO

FDPM

FFIDP

GF

GI

G. Vol.

ha

IPM

IUFRO

:r..cx;PROO

LP

LPPM

LRSY

MEM

MF

MIDA

MR

NFl II

ASEAN Institute of Forest Management

cubic meter

Centimeter

: Diameter Breast Height

Food and Agriculture Organization

Forestry Department Peninsular Malaysia

Forestry and Forest Industries Development Project

Good Forest

Growth Index

Gross Volume

Hectare

International Business Machine

International Union of Forest Research Organization

Log Production Planning Model

Linear Prograrrrning

Log Production Planning Method

: Long Range Sustainable Yield

�Blaysian Economic Model

r-bderate Forest

Malaysian Industrial Development Authority

: Malaysian Ringgit

National Forest Inventory II

ix

PF : Poor Forest

PFE Permenent Forest Estate

SF Superior Forest

SLF : Stateland Forest

SR : Surv ival Rate

STANDPRUJ: Stand Proj ection r-trlel

sq . m

UNIOO

YEAr..

: Square meter

Un ited Nat ions Industrial Development Organizat ion

Year Elapsed After Logging

x

Abstract of thesis submitted to the Senate of Universiti Pertanian t-1alaysia in partial fulfilment of the requirements for the degree of Master of SCience

STATE LE.VEL roo PROa.JCl'IOO PLANNING USING LINEAR PR('X;RAMMING

By

ABOOL RAHMAN ABIXJL RAHIM

JUNE 1989

SUpervisor Yusuf Hadi, Ph.D.

Co-SUpervisor Awang Noor Al:rlul Ghani, �1. Sc.

Faculty Forestry

This study aims to develop a log production planning

nodel and to apply it at State-level planning. In the near

future, t-i:l.laysia will have to rely rrore and rrore on the

Permanent Forest Estate (PFE) , after the State land forests have

all been converted for agricultural development. A more

efficient and effective log production planning model is

essential to ensure that the forestry sector continues to

contribute to the socio-econanic development of the country.

A � E.roduction flanning t,.1alel (LPPH) was developed,

consisting of a stand proj ection ( STANDPRQJ) model and a Linear

Programning (LP) (lDGPROD) model. Data for the models were

xi

collected fran the State of Negeri Sembilan. STANDPROJ

estimates forest yield which are used, together with forest

areas am forest rranagement p::>licies , to generate an LP

problem matrix in LOIUS spreadsheet . The matrix is accessed

am solved by an LP optimizer . 'll1e optimizer also transmi ts

the solution back to LOIUS which acts both as a matrix

generator and rep::>rt writer.

LPPM forecasted that the annual log prcrluction for Negeri

Sembilan v.ould decline fran 660 , 000 cu. m in 1 980-1 989 pericrl

to 460 , 000 cu. m in pericrl 1 ( 1 990-2000 ) , then declining at 20

percent per decade until pericrl 4 . Fran pericrls 5 to 7 ,

production fluctuates ..:t.1 0 percent arourrl a long f.ange

sustainable yield (LRSY) of 21 0 , 000 million cu. m per year .

Inclusion of an additional cutting cycle of 40 years increased

log prcrluction. Additional information fran the LP solution

includes the relative importance of the forest types and the

lower and upper limit of their sizes for the LP solution to

remain valid.

LPPM provides a fast and efficient methcrl of analysing

log prcrluction and is able to prcrluce a !TOre favorable trend of

log prcrluction. 'll1e m:rlel is recarmended for implementation at

state level . The forecasted log prcrluction for Negeri Sembilan

shows a declining trend, and additional supply of log

prcrluction needs to be obtained from forest and rubber

xii

plantations and forests in other states to cater for the State

varl-based iooustry.

xiii

Abstrak tesis yang dikemukakan kepada Senat Universiti Pertanian Malaysia sebagai memenuhi sebahagian daripada syarat keperluan ijazah Master Sains

� PENGELUARAN KAYU BALAK DI l'ERn-G<AT NEXiERI MEN:nJNAI<AN PROORAMAN LINEAR

Oleh

JUN 1 989

Penyelia Yusuf Hadi , Ph.D.

Penyelia Bersama : Awang Noor Al::rlul Ghani, M.Sc.

Fakulti Perhutanan

Kajian ini bertujuan untuk IreI1yediakan satu rocrlel

perancangan pengeluaran kayu balak bagi digunakan di peringkat

Negeri . Di masa a.kan datang pengeluaran kayu balak akan

bergantung kepada kawasan Hutan Simpanan Kekal bila Hutan Tanah

Kerajaan habis dibangunkan untuk pertanian. Satu rocrlel

perancangan pengeluaran kayu balak yang cekap dan berkesan

diperlukan supaya sektor perhutanan terus IreI1yumbangkan kepada

pernb:mgunan sosio-ekonani negara .

Satu M:Xlel Perancangan E.engeluaran !5.ayu �lak (MPPKB)

telah diajukan , IreI1gandungi M:Xlel Unjuran Dirian Hutan dan

M:Jdel Perancangan Pengeluaran Kayu Balak Menggunakan Programan

Linear (LP) • Maklumat-mak.lumat bagi rrcx:lel-rncx1e1 tersebut

xiv

diperolehi dari Negeri Sembilan. M:rlel Unjuran Dirian Hutan

menganggarkan hasil dirian hutan. Hasil ini digunakan bersarna

keluasan hutan dan p:>lisi p:mgurusan hutan untuk membentuk

matriks masalah LP menggunakan program WIUS. Masalah ini

selanj utnya diselesaikan dengan rrenggunakan 'optimizer' LP yang

juga memasukkan hasil p:myelesaian ke dalam WIUS. Program

WIUS bertirrlak sebagai 'Penyedia Matriks' dan juga 'Penulis

Laporan' •

MPPKB rnengunj urkan pengeluaran kayu !:alak !:agi Negeri

Sembilan akan menurun dari 660 ,000 meter padu setahun !:agi

tempoh 1980-1989 kepada 460 ,000 meter padu setah\.ll1 regi tempoh

1 (1990-2000) , dan seterusnya menurun pada kadar 20 peratus

sedekad sehingga tempoh 4. Dari tempoh 5 hingga 7 , pengeluaran

beruba..h-ubah dalam lingkungan kadar .±.10 peratus dari PRJ

(Pengeluaran Berkekalan Jangkapanjang) sebanyak 210 ,000 rneter

padu setahun. Penggunaan p.1singan tebangan tambahan 40 tahun

meningkatkan lagi p:mgeluaran. Maklumat lain yang diperolehi

dari MPPKB ialah perbandingan kepentingan jenis-jenis hutan dan

had keluasan hutan yang boleh digunakan supaya penyelesaian LP

tidak berubah.

MJdel yang diaj ukan ini l::nleh rrenganalisis pengeluaran

balak. dengan cepat dan cekap, dan menghasilkan tren pengeluaran

!:alak yang lebih sesuai. M:rlel ini disyorkan supaya

dilaksanakan di peringkat negeri. Unjuran pengeluaran kayu

balak yang rrenurun rrenunj ukkan ba.hawa Negeri Sembilan per lu

x.v

rnencari sumber lain seperti ladang hutan dan getah serta kayu

balak dari negeri lain bagi menampung keperluan industri

berasaskan kayu-kayan di negeri tersebut .

xvi

Problem statement

Log production planning involves the determination of the

timing and extent of forest areas to be harvested and

subsequently to regenerate in order to ensure acceptable log

production in the short term and sustainable level in the long

run. Log production planning is seen as a part of forest

sector planning and involves an analysis of the resource

supply_ The other part of forest sector planning involves an

analysis of the utilization, demand or requirements for the

forest resource (FAG , 1974a) . Both parts are essential for the

sound develop:nent of the forestry sector. other terms used to

describe log production planning include Timber Management

Planning, Harvest Scheduling, Timber Harvest Scheduling, Forest

Regulation and Timber Supply Planning and Log Supply Planning.

Malaysia, as a rrajor supplier of tropical timber, needs an

efficient and effective method of log production planning.

This is important so that the forestry sector will always

continue to contribute to national developnent.

Planning is necessary to ensure that forestry rraintains

its contribution to socio-econanic develop:nent and

1

2

envirorunental protection. The forestry sector is one of the

rrajor sectors of the Malaysian econany. It contributes to

foreign exchange earnings, employment, state revenues, value

added and regional development. The sector also offers

intangible contribution such as the rraintenance of

envirorunental stability, minimizing darrage to rivers and

agricultural land by floods and soil erosion and the

safeguarding of water supplies and conservation of genetic

resources. In Malaysia, the major forestry prooucts which

include sawlogs, sawntirnber, pl�, veneer and Iroulding have

shotm an increasing contriootion to the country exr.:ort value

fran MR 4. 7 billion in 1 986 to about MR 6 . 8 billion in 1 987 .

The sector has also provided employment to about 1 46 , 000 people

in 1 987 , accounting for about 3 percent of the country I s total

employment for that year. In Peninsular Malaysia, the forestry

sector has contriooted about MR 1 . 8 billion of foreign exchange

earnings and employment for 7 4 , 51 7 people in 1 987 . In 1 988 ,

the sector provided employment for 89 , 347 people and generated

about MR 2 . 1 billion of foreign exchange earnings.

The planning methoo needs to re improved to plan log

proouction frc::m a gradually shrinking resource base. Over the

years logs were proouced frc::m roth the Perrranent Forest Estates

(PFE) and state land Forests (SLF) in Peninsular Malaysia. In

the future, logs will cane rrainly frau PFE, as the SLF are .pa.

being developed for agriculture and other uses. The Forest

Department Peninsular Malaysia forecasted that the reloggable

3

SLF of 0 . 91 million hectares in 1 984 YA:)Uld all be harvested by

the year 1 995 , based en the planned oonversion of 80 , 939

hectares per year (Thang, 1 985 , 1 986) .

Fran 1 986 to 1 993 , Peninsular M:ilaysia is forecasted to

prcxiuce 8.06 million cu. m of logs per year. 'Ihe log

prcxiuction is expected to decline to 5 . 99 million cu. m in 1 994

when all SLF are expected to have been canpletely harvested

('lllang, 1 985 , 1 986) . Efficient planning nethcxi is thus needed

in light of this forecasted declining log prcxiuction.

The current practice of log prcxiuction planning in

Peninsular Malaysia employs manual area-control calculations.

It does not incorporate detailed yield estimations for the

respective forest types. Forests are grouped into areas

managed under 30 and 55 years cutting cycle with current as

¥/ell as future yield of 61 . 23 and 77 .00 cu . m per ha,

respectively.

A IOCrlel is proposed that would provide sane improvements

to the current practice. It incorporates a stand projection

IOCrlel to refine estimates of future timber yield fran the

forests. '!his canponent of input data has been identified as

being the weakest link in past planning IOCrlels (Yusuf, 1982) •

It intrcxiuces an intermediate cutting cycle of 40 years in

addition to 30 and 50 years to provide flexibility in forest

managerrent. Linear Prograrnning (LP) is used to provide for a

quick and efficient assessment of log prcxiuction level. It can

4

be used to conduct a sensitivity analysis of future log

pr oduction. '!he m:x1el is r ecarmerrled to be applied to state

pl anning situation to reflect the real situation in Peninsular

Malaysia , where the twelve state Governments have the

juri sdiction over the forest resources .

z.bdels such as the me proposed in this study are able to

regulate the forest in order to achieve sustainable log

production vol urce. It w:>uld be useful if the m:x1el can

prioriti ze harvests fran the forest , for e xample , the • oldest

sta nd' first or trees fran t he best sites first. IbNever ,

prioriti zing of the harvests fran the forest is not possible at

t he moment in Malaysia because t he inf ODmation on age classes

and site index are not readily a vailable .

LP is chosen as a techni que because it is an efficient and

quic k way of analysing log prcrluction level . It is able to

capture the greater rart of problem of l OCJ prcrluction planning

and came up with an optimal solution for a given set of

ass umptions . '!he problem of all ocating forest areas to several

alt ernative management r egi mes is a typical probl em solved by

LP. Several runs of the m:x1el with different assumptions would

help the planner in making decisions . LP ' s capabil ity is

indicated by its wide appli cation for l OCJ prcrluction pla nning

in North America , Europe , Ja pan, Australia and Malaysia .

Log prcrluction m:x1els employing Linear Progr amming

include MAXMILLIAN ( Clutter , 1 968) , Timber RAM ( Navon , 1 971 ) ,

5

FORPLAN (Joh nson and Jones , 1979) and SAM (Field, Wilson and

Fl owers , 1980). Variation of these models have been a pplied in

Canada ( Erri co, 1981), Japan ( Konohira , 1981), Ollie ( Barros

a nd Weintraub, 1980) and Taiwan ( Yang, Lin, Chen a nd Wang

1980). Malaysian applications have been descri bed by Yusuf

( 1982, 1987); Tieh, 1986); Yusuf , AlXl. Rahman and Awang Noor

( 1988); O1ai ( 1988) and Lim, Ferrie a nd Si m ( 1988). LP rncx:lels

have been developed in Malaysia for other forestry purposes .

'Ihese include log transportation rncx:lels (FAD, 1975; Sandum,

1988; Liew, 1988; Yusuf , Shukri and Vincent , 1988), and mill

l ocation model ( Kharnurudin a nd Tonga , 1988).

Objective of the Stooy

This study ai ms to develop a pla nning model incorporating

Linear Prograrrrning and sta nd Projection to analyse log

pr oduction, and to illustrate the operation of the rncx:lel by

applying it to plan log production for the State of Negeri

Sembilan.

Scope of the stooy

In this st udy , the number of ha of forest land to be

rranaged by al ternati ve management regi mes is dete rmined l::ased

on maximi zation of total log volume proouction over a 70-year

planning hori zon. This physical vol ume pr oduction criterion

was used instead of discounted revenues , profits , present net

values or oosts tecause of lac k of data 00 the econanics of

forest operations in Malaysia.

6

The data used in the st udy are fran Negeri S embilan. '!he

results , therefore , pertain only to t hat state . However , data

f ran other states could be used in the model to analyse their

log production.

Lack of detailed data pennitted the forests to be

classified into only four productivity classes as in the

National Forest Inventory of Peni nsular Malaysia. Data on

gr owth and rrortality of trees �e equally scanty. '!he si mple

nature of the starrl proj ection model is because of this.

Both LP and LOIUS spr eadsheet prograrnnes are ca pable of

harrlling rrore canpl ex problems . Such problems could be

constructed after rrore detailed data are a vailable.

Organization of the Report

'!he thesis is divided into six c hapters . Chapter II , the

review of li terature , sunmarises previous v.orks in log

pr oduction planning using Linear Progr amning (LP) . Methods of

log production planning and log production planning using LP

are highlighted. Chapter III descri bes the methodology of

study. Chapter IV presents the result of the study. It

descri bes the structure and operation of the pla nning model ,

log production pattern for Negeri S embilan and the sensitivity

of forecasted log production for the state. Chapter V

discusses s pecia l features of the model, comparison of

forecasted log production, areas for further res earch and a nd

7

i mplications of t he forecasted log production. Chapter VI

hi ghli ghts the conclusions and r ecarmerrlations.

ClIAPlER II

RE.Vnli OF LITERA'IURE

Methods of Log Production Planning

Meth ods for log pr oduction plarming could be classified

into t\\1O major groups: traditi onal and the modern methods

( Yusuf , 1 988) • 'llie traditi onal methods incl ude area control ,

volume control and canbination of area-vol tnne control.

Area Cootrol

'Ihis approach r equires cutting equal areas annually or

peri odically. The a nnual harvest area or a nnual coupe is the

total number of hectare of pr oductive forests divided by the

cutting cycle or rotation ( Leuschner , 1984; Davies and Johnson,

1987) . 'llie advantages of this approach are its s �licity ,

ease of identification of the area on the ground, and its

pr oduction of a regulated forest. Its di sadvantages are the

l arge fluctuation in the level of log pr oduction and the need

to canbine it with sane ty pe of vol ume control for a rrore

meaningful planning exercise .

This appr oach is widely used in Malaysia. Using this

a pproach, log su pply fran Penuanent Forest Estate in Peninsular

Mal aysia is pl armed by classifying the forest into two

a:mpooents and then a pplying the correspoIrling cutting cycle

8