A quantitative model for the design of nutritious and acceptable foods : a thesis presented in partial fulfilment of the requirements for the degree of Ph. D. in Product Development at Massey University

(1)

Copyright is owned by the Author of the thesis. Permission is given for

a copy to be downloaded by an individual for the purpose of research and

private study only. The thesis may not be reproduced elsewhere without

the permission of the Author.

(2)

A QUANTITA T I VE MODEL FOR THE DES I GN OF NUTR IT IOUS AND ACCEPTABLE FOODS

A thes is pre s ent e d in partial fulfilment of the requirements for the degree ^{of Ph.D.}

in Prorlu c t Deve l opment a t Mas s e y University

Allan Malcolm Anderson 1975

(3)

i i

ABSTR�CT

A quant ita tive model was deve lope d for th e s ys te ma tic s e l e c t ion of raw ma t e r ia ls for nu tritious and a c c e pt a b le Tha i f o o ds . The bas is for the quantita t iv e mode l was l inear

pr o gramming . Unlike m o s t o th e r a pplications of this t e c hnique which c ons idered only nu tritional ne e ds , th is mode]. a ls o took int o acc oun t th e c onsumer ' s r equ irement s f or an a c c e p table f oo d . The d e v e lopment of the �o d e l formed part of a r e s earch pro gramme in the Food Tcchcology Department o f Mas s e y University whi c h is inv e s tiga t ing quan t i t a t i ve p r o du c t developmen t t e ch niques.

The linear program�ing model was used f irst ly to select

raw materia l mixes wh i c h w ould sa t is fy the da ily nutr i t i onal requ ir e ments of the Tha i pecple . S e le ction was ^made f r om a lis t of 1 5 1 availa b le raw materials ^- 144 �ndigenous Thai raw materials

and 7 New Z e a land da iry produ c ts . The requirements for 26 nutr ients w e r e sa t is f ie d . The s e inc lu de d pr o te in , fa t , ca lorie s , f ibre ,

3 ^miⁿer^als , ⁹ vitamins and 1 0 e s sential amino a c ids . The lack of

sp e c i fic upper limits on most of the n u tr ie nt s resulte d in

s o lutions w i th a gross nu tritiona l imba lance . Rec ent investigations

indicate that such imbalances may be de trimental to human h e a lth and i t is sugge s t e d th a t a more s a t is fa c tory diet ie one where a ll nutr ie nts are ba lanc e d , a t or near the ir l ow e r re qu irement leve l s . C ons i derable �r oblems exis t in a c h i e v ing a ba lanc e d die t using range c ons traints in the linear programme �ae to the probab i lity of s o lut ion infeas ib i l i t ie s .

Goa l programming , an e xt e ns ion o f l j near pr ogramming , ha s b e e n u s e d in oth e r f ie lds o f r e s earch t o minimize t he d e v ia t ion of s o l u t ion var iab le s from sp e cⁱfic goals . This t e c hnique show e d p o t e n t ia l in a t ta ining a balanc e d nut r i t io na l d i e t wher e the goals

represented the requ irements for spe c if ic nutr ie n t s . A goal

programming mode l was ^devised whi ch f ir s t ly a ch ieve d a ba lance of e s s e n tial amino a c i ds a s c lo s e to tha t of e gg pro t e in a s pos s ible . S^ec on^dly , a so lution was o b ta in e d where a l l

26

nut r ients w e r e a t

(4)

I

i i i

the ' op t i mu m ' balanc e . ^The a c h i e v ement o f a bala nc e d nut r i t i ona l

die t r e s u l t e d in a large i nc r e a s e in cos t and indica t e d th e impor tanc e o f care ful de finit i on o f the r e quⁱrements f o r

nutr i t i ona l balance in future r e s earch whe re c o s t minimiza tion is a pr i or i t y .

The raw ma t e r ia ls s e le c t e d by b oth goal nnd l inear

programming were t o ta lly una c c e ptable as ingre d ien t s ⁱn a Tha i f ood dish w i thout ext e ns i ve pr o c e s s ing to cha nge both flavou r and texture o f the m ixture . It was mor e logic al to prov ide a p r o c e dur e for raw ma ter i a l s e le c t ion on t he bas is of the ir c ombine d a c c e ptab i lⁱt y

in a Tha i fo od dish . Nonme tri c mu lt i d imens iona l s c a ling was u s e d t o der iv e a 3 d imen siona l c on f iguration of ⁴⁰ raw ma t e r ials fr om c on sumer inf orma tion on the us e o f thes e raw mat e r ia ls in Tha i f o o d d ishe s . The a x e s of thi s s pac e r e presente d the domina n t pr ore r t i e s o f r a w ma t e r ials in determining f o od dish a c c e p ta bi l i t y . An ide a l p o int w a s located in th is spa c e . This p o int was de f in e d as the

' opti mum' c o mbinat ion of raw ma t e r ial pr ope r t ies required in a Thai f o o d d ish .

Nonme tr ic mu ltidimens i o nal scali�g provi d e d th e bas is for der iva tion o f me tr i c s ca le va lue s for t he ⁴⁰raw mate r ia l s and the ideal point^. The s e valu e s w e r e u s e d to derive ^a l inear fu�c t ion r e la t ing r a w ma terials to f o o d d i s h a c c eptability. This fun c t ion was use d i n the linear programming model toge ther with nutr it j_ ona l c ons tra ints to provide a sys tema t i c me thod o f raw mat e r ial s e le c t i on for nu tr i t ious and a c c e ptable Tha i food dishes . The raw ma t e r ial mixes s e le c ted by th is mode l showe d a marked ⁱmpr o v e m e nt ov e r those ch osen by the linear programme s ub j e c t to only nutri tional

c on s tra int s .

(5)

iv

ACKNOWLEDGEMENTS

The re search c o ntaine d in this thes is was supporte d by

f inanc ial a s s is tance from a f e l lows h i p provided by Rank Xerox ( NZ ) L t d . The al loc a t io n of t he fu nds from t h is f e llowBh i p were a dm inj sie r e d by the New Z ea land Da iry R e s earch I ns t i tu t e . O th e r f inancia l a s s istance w a s o b t a i n e d from a Mau r i Br others and Thomson s ch o larsh ip.

I will al^ways be gra t e ful to my main superv is or� Dr Mar y Earle , f or the confidence , enc ourageme n t an d above a ll the

patience she sh owed dur in g a l l s tage s o f the pr o jec t . To my

s e c ond supe r v is or� Dr B i l l Schroder! I express my appre c ia tion for h is man y us e ful sugges t ions and c r i t i c isms .

The suc c e s s ful c omple t ion of the res earch was largely due

to the cooperati on obtained from Chulalongkcrn Univcrci ty ⁱⁿ

Bangkok . I n par tic ular I wou l d l ike to thank Patchre e fo1· making my v is i t to Tha iland poss i b le , for h e r h e lp in translating

que s t ionna ire s and f or t e a c h ing me so muc h about Thai customs.

I am als o indebted t o Cha iyute for his help an d gui da nce duri11g my stay in Ba ngkok and t o Kum j on and Yuthana for the ir a s s i s tanc e w i th the c o nsumer surveys .

'I w ould als o like t o e xpr ess my appre c ia t io n t o many

pe ople , both in ^New Z e a land and overs eas , who a s s i s ted i n a varie t y o f ways . I n par t icu lar m y thanks g o t o :

Dr B i ll Edwardson for ini t ia t ing the research p r o gramme o f which this the s is f orms a par t .

Dr C olin Bosw e l l and the s ta ff o f the C ompu t e r Unit , Mas s e y Univers ity , f o r the c ompu t ing fa c iiities and pro gramming a dv ic e . The library s t a ff a t Mass e y University for the ir w i lling

ass is tanc e in locating r e f e r e n c e ma t e r ia l .

P r o f e s s or Richar ds and the Food T e c hnolo gy Depa r tment for the use of facilit ien and for the ir s uppor t throughout .

(6)

V

Dr Ian S p e nc e , Univers i ty o! W e s tern Ontario and Dr J ohn Ogilvie , Univers ity o f Toronto , for the ir advice o n the applicat ion o f mul t idime n s i ona l s c a l ing .

J ane t M ilne for eve n tua lly agr e e ing t o type the manu s c r ip t and for mak ing s u c h a gre a t j ob of i t .

Fina l ly , to my w i fe J a ne t , who has shar e d in a ll the ups and downs o f the pas t thr e e year s . Her pos i t ive c on t r ibu t i ons t o the suc c e s s ful c omple t io n o f th i s the s is are t o o nume rous t o

ment ion .

A llan Anderson Palmer s t o n N o r th New Zealand De c e mter , ^{1 9?5}

(7)

1 ^•

1 • 1

1. 2

1. 3

2.

2.1 2.2 2.3 2.4 2

.

⁵

2

.

⁶

? '7

� • I

3. 1 3.2 3·3 3.4

7. c:;

./ • ./

4.

4. 1 4.2 4.4 4.3 4.5 4.6 4.7

5·

5. 1 5.2 5.3 .5 .4

CONTENTS

'rABLE^S

FIGURES PREFACE

THE ENVIRONMENT AND THE MODEL The env ironme n t

Th e mode l

St..,nrnary

LINEAR PROGRAM�ING AND NUTRITION

A brief r e vie w of linear programmⁱng applica t i ons in the fo od ^industry

Th e data r e quir�d The raw mater ials

Nutritional constraints

The w e igh ted âverage ^per capita da i ly r e quirements S e tting u p the da ta for c ompûter so^lutiôn

Sumrnar:l

SOLUTIONS TO THE NUTRITION PROBLEM The l inear programrr.ing approa c h_

)

The applica t ion of goal programming Goal programm ing on the amin o acids

Goal programming on all nu tr ients Summary and c onc lusi ons

FOOD CHARACTERISTICS AND C ONSUMER ACCEPTABILITY What is consumer a c c eptanc e ?

The c las s ification o f s ens or y propertie s o f f o o d s T h e measurem e nt of the s e n s ory proper^tⁱes o f ^foods The measurement o f c o ns umer a c c e p ta n.:·� of fo o d The pr ob lems i n u s i ng c onventiona l sens ory me th ods A ^Etudy of multivar iate te chniqu e s

S u mm^ary and c onc lus iors

MULT IDIMENSIONAL SCALING - THEORY AND APPLICATIONS What ^is multid imens ional s c a ling ?

Nonmetr ic multidimensiona l s ca l ing

Applications of mu ltidimens ional ^scaling Summary and c onc lus ions

v i

v i ii

X

xi

18

21 22, 28 Lf8 50 55

56 56 60 64 69 75

78 79 85

104 96

109 112 114 116 117 1 19

128 132

(8)

5.

6.1 6.2 6.3 6.4 6.5 6.6 6.7

7- 7. 1

7.2 7.3 7 • L�

7-5 7-6 7-7 7. 8 7-9

8.

8. 1 8.2

8.3 8.4 8.5 8.6 8.7

1 •

2.

3.

4.

5.

6.

DATA C OLLECT I O N FOR MULTI DIMENS IONAL SCALING ANALYS IS Ge neral c omme n ts on data c olle c t io n

R a w ma t e r ia l s imila r i ty data Raw ma terial a s s o c ia t ion da ta

' Raw ma terial by u s e ' da t a Food d i s h s i m ilar ity data

Raw ma t e r ia l and food dish a c c e ptab ility S ummary and c onc lus ions

MULT I DIM ENS I ONA L SCALING ANALYSES

The applic a t i o n of mu l t i d ime n s ional scal ing Raw ma t e r ia l s imilar i t ie s analys is

Raw ma teria l asso c ia t ion analysis ' Raw ma terial by use ' ana lys is C ompar ison o f raw ma teria l spa c e s Fo o d dish s i m ilar i t i e s analys is

Unfo lding analysis on the raw mat e rials and f o o d di.sh�s I deal point loc a t io n

S ummary a n d c on c lu s i ons

I NCLUS ION OF ACCEPTABI LITY FAC TORS IN THE ^{L. P .} MODEL The der iva tion of a linear function of a cceptab ility Data or gan i z a t ion for a c c eptability opt imiza t i on u s ing L . P .

L inear programming solut ions A s ummary o f the t o ta l model Valida t ion of the mode]

The flexib i l i t y of the mode l S umm�ry and c onc lus i o ns

QUANTITATIVE PRODUCT DES I GN I N PERSPECTIVE BIBL IOGRIIPHY

APPENDICES

Raw ma teria l c ompos i t ion ta b l e s Da ily nutr i t i onal r e quir e me n t s

R a w ma terial s use d in the nut r i t i o n problem Dis tr ibut\o n of the Tha i popu la t io n ⁽¹⁹⁷⁰⁾ S o lu t io�s t� the linear and goal pr olramming applica tiuns t o th e nutrit i o n pr oblem

The da ta for mult i dimens i o na l s c a l in g ana lys is De s c r iption o f 20 food dishe s us e d in the food dish s�m ilar ity s urvey

8. L . P . s olu t ions w i th nu t r i t ional and a c c e p ta b i l i ty c ons traints

v i i

135 136 142 143 145 14 8 150 152

154 155 158 164 16 8 172 173 17 8 1 81 183

1 85 1 86

1 89 191 195 196 205

206 208

213

230

240

242

244

246 251 257 260

(9)

1 ^•1 1 . 2

1 . 4 1 . 5 2 . 1

2 . 2 2 . 3 2 . 4 2 . 5

2 . 6 2 . 7 2 . 8 2. 9 2 . 1 0 2 . 11 2 . 1 2 2 . 1 3 2 . 1 4

TABLES

O c c urre nce o f nu t r i t ional de fic iency among pat ie n t s a t te nd ing 6 2 h o s p itals in 57 provinc e s o f Tha i land

( 1 9 70 )

Cumu la t ive p e r c e n t a ge o f h ous eholds having inc o me l e s s than a spe c if i c c la s s , mun ic ipa l area and v il lage P e r c ent expenditure per household f or c onsumer goods and s e rvices

Household expe n d i ture . P e r c e n t d is tribu t ion o f expenditure on var i ous f o o d groups

Quant i t ie s of f o o d s c o n&ume d by food groups ( 1 962 -63) Summary of l inear pr ogramming mo deJs in nu t r i t io n prob lems

The nu trients c onsidered in the. proble m S ourc e s o f raw ma t e r ial c ompo s i ti o ns

W e i gh te d average da ily per capita nutr i t i ona l r e qu ir e ments

Re c omme nde d daily calorie allowanc e s for infants ch ild r e n

Re c o mmende d da ily pr o t e i n allow&Hces C a lor i e s from fat in Eastern c ountries R e c omme nde d da ily c a l c iu m allowanc e s R e c ommended daily a llowanc e s for ir on Re c omme nde d daily a llowan c e s for v itamin c

R e c omme nde d dai ly allowa n c e s for v i tamin A R e c omme nde d d a ily a llowanc e s for v i tamin B 1 2 R e c omme nde d da ily a llowan c e s for v i tamin B6 R e c omme nde d dai ly a llowanc e s f o r ' fr e e ' f o late

and

v i i i

4

6 7 8 8

20 2 4 2 5 2 9

32 33 34 36 38 39 40 4 3 44 45 2 . 1 5 R e c om�e nde d da i ly a llowa n c e s for pan t o th e nic a c id 46

2 . 1 6 S e l e c t e d e s s e n t ia l amino a c id pa t t e rns 4 7 2 . 1 7 Nut r ie nt s r e quir ing e xtra a llowanc e s dur i ng pr e grian c y

and la c ta t ion 49

3.1 P e r c e n t de via t i on of nut r ients from goal pr o gra�me on

ind ige nous raw ma t e r ials 72

3.2 P e r c e n t devia t ion o f nutrients from gcal pr o gramme

us ing all raw ma t e r ia l s 73

4 . 1 A h i s t ory o f tas t e c la s s i fic a t ion ( Boring ( 1 3 ) ) 86 4 . 2 A h is tory o f o dour c la s s i ficat ion ( Boring ( 1 3 ) ) ⁸⁸ 4 . 3 A s y s t e m o f t extur e c las s i ficat ion ( S zc z e sniak ( 1 98 ) ) 93

..

(10)

ix

4.4 C ompar ison o f the s u ita b i l ity o f various multi var iate

t e c hnique s f or the s e c o nd stage of the r e s e a r c h 1 1 3 6 . 1 Raw mate r ials us e d in the multi dimens iona l s c a l ing

ana ly s e s

6 . 2 The 1 5 Tha i f o o d d i s h c ate gor ies u s e d in the ' raw mate r ia l by us e ' sur v e y

6 . 3 Average leve l o f u s e in f o o d dishe s for 4 0 _raw mater ials

6.4 _The 20 food dishe s u s e d in the food dish s im i la r ity surv e y

6 . 5 6 . 6 7. 1 7 . 2 7·3 7 . 4 8. 1

8 . 2 8 . 3

Rank order o f raw mate r ia l a c c e ptability Rank order o f food dish a c c e ptabi l ity

S c a l e va lue s for the raw mate r ial similar ity spac e S c ale valu e s for the raw mater ia l assoc iation s pa c e S ca l e value s for the ' raw mater ial b y us e ' spa c e S ca l e valu e s for the f o o d dish spa c e

W e i gh t e d average daily nutr it iona l requ ir e m e nts f or 20-29 year old T�a is

S o lut ions from the parametr i c runs optimi z i ng ACCEPT S o lut i ons to parametr ic changes in PREFAC w ith

C ONPFAC at z e r o

8.4 S olu t i ons t o parametric c ha�ges in COMPFAC with

PREFAC at 1 0 . 6 79

1 3 7 1 46 't48

1 49 1 5 1 1 52 1 59 1 65 1 69 1 75

197

1 98 8 . 5 Raw materia l c omb inations us e d in Thai taste pane l s 201 8 . 6 The actual and predicte d a c c e ptabi l itie s for 21 raw

mate r ia l mixe s 202

..

(11)

2 . 1 2 . 2 3 . 1 3 . 2 3 . 3 4 . 1

4 . 2 4 . 3 4 . 4

7 . 2 7 . 3 7 . 4 7 . 5

FIGURES

L inear pr ogramming matrix f or the Thai nu t r it ion problem

Row and c o lumn summa ry f or nu t r i t io n problem . Lis t e d b y th e SUMMARY pro c e dure LPS/1 1 30

E f f e c t of impr ovement in amino a c i d ba lan c e on c os t Graphical r e pr e s e n ta t ion o f a n L . P . problem

C ar d input for o b j e c t i v e func t ion of goal programme on amino a c ids

Par t o f a c onsumer beha vioura l mod e l ( Engel e t a l ( 43 , p46 ) )

The tas t e t e trah e dron ( H e nning ( 8 1 ) ) The smell pr ism ( He nning ( 80 ) )

A pro c e dure for e valuat ing t e x ture ( S z c zesniak and Kleyn ( 201 ) )

The 9 -point H e d on ic S ca le

KYS T c ontrcl c a r d s for s c aling s olut ions in 5 through t o 1 dimens i ons

P lo t s for the raw ma t e r ia l s im i la r i ty spa c e P l o t s for t h e r a w ma terial a s so c ia t io n spa c e P l o t s f or the ' raw ma t e r ial b y use ' spa c e P l o t s for t h e f o o d dish space

7 . 6 KYS T contro l car ds for the unf olding ana lysis o f raw

X

5 1 52 6 3

8 1 8 6 89

95 1 06

157 1 60 1 66 1 70 1 76

ma t e r ia ls and f o o d dis h e s 1 79

7 . 7 KYS T control c a r ds f or the id e a l po int lo c a t ion 8 . 1 P l o t o f PREFAC a gains t COMPFAC for 21 raw ma terial

c omb ina t ions

1 82 2 0 4

(12)

xi

PREFACE

Food product development is the procedure by which raw materials are selected and then combined through a system of

processing into a food product acceptable to the final consumer. For many years, product development has been based more in intuition than on objective judgement. There has, however, been a gradual infiltration of quantitative techniques in recent years in an attempt

to brin g greater objectivity into the system of product development.

Some examples of tWese techniques are:

Idea generation by systematic methods

Market evaluation based on Bayesian decision tree methods or cost benefit analysis

Formulation based on linear programming methods Business analysis including market forecasts, market

surveys and buyer simulation models.

The introduction of these techniques has been largely in a haphazard fashion. The Food Technology Department at Massey University has begun a research programme to develop a systematic approach to the application of quantitative techniques through all stages of product development. The design of nutritional products for consumers in developing countries is being used as the basis for this programme.

An overview of the entire research programme is given in a thesis by Edwardson (41).

Edwardson provided an objective procedure for the selection of raw materials to be used in low cost nutritional products for the Philippines. The major problem encountered by Edwardson was the

incompatibility of raw materials selected by linear programming based on nutritional criteria. It was only possible to produce an acceptable product from these raw materials through extensive processing and addition of flavouring ingredients.

It was the object of this thesis to both complement and extend Edwardson' s work w�th the application o.f the linear programming model to the selection of raw materials for the production of low cost food dishes

(13)

xii

which would meet the nutritional needs of people in Thailand. Emphasis was placed on the extension of Edwardson's model to include consumer acceptance criteria. A procedure was devised whereby raw material

selection was made on the basis of their predicted acceptability in Thai food dishes.

Thailand was chosen as the environment for the study in preference to the Philippines because a cooperative programme had recently begun between the Department of Chemical Technology at

Chulalongkorn University in Bangkok and the Food Technology Department at Massey University. Information on Thailand was therefore more readily available.

Thailand has a reasonable supply of food raw materials and even exports rice and sugar, but problems of malnutrition still exist, particularly in the North and North-east regions of the country. There is a definite need for the locally grown raw materials to be put to the best use through their optimum combination in acceptable food dishes.

ln this thesis, considerable emphasis was placed on the selection and use of indigenous raw materials.

Supplementation of local materials with imported foods has played an important role in the design of nutritional food product�.

Included among the most commonly used supplements are dairy products.

The use of some dairy products as ingredients in local Thai dishes was evaluated both as complements to indigenous raw materials in the supply of essential nutrients and on their general acceptability in these dishes.

This thesis first reviews the availability of foods in Thailand and then suggests how the linear programming model could be used to

design nutritional foods. The nutritional compositional data for a wide variety of indigenous raw materials and imported dairy products were

collected, and the per capita daily requirements of the different nutrients in Thailand were found. Using this information in the linear programming

(14)

xii.i

mode l , raw material mixes were s elected to m eet the nutritional needs of the general Thai population .

Goal programming , which is basically an extension of linea r p rogramming, wa s applied in the selection o f raw mate ria l mixes t o give nutritionally balanced diets . This technique was a l so used to evalua te dairy products as complements to indigenous raw materials in p roviding firstly an ideal balanc e o f essential amino acids and se condly a balance of all 26 nutrients which were considered in the model.

A procedure for raw materia l se lection based on their

acceptability in combination in food di shes was obviously required in the 1 inear progrannning model . A reviet., of the literature relating to sensory evaluation and consumer a c ceptance o f f oods f ailed to provide any procedure which might be used directly. The problems of de f inition and mea surement of the raw materta l prop erties wh ich influence the acceptability of ^fôôd dishes d_irec_ted the research tôward a study of multivaria t e techniques. Nonmetric multidimensional scaling was i d entified as a technique which could not only resolve the properties

but could also provide metric scale values for these p roperties which might b e used directly in the linear progrannning model.

A subset o f the original list of raw materials wa s selected.

Proximities data on these raw materia ls were obta i ned from consumer surveys in Tha i l and and were subj ected to multidimensional scaling analysis. A J dimensiona l s patial configuration o f the raw materials wa s derived and the axe s of this space were identified as the dominant p rope rties o f the raw materia l s in dete rmining their acc eptability in Thai food dishes . A linear function was derived re lating these raw mate rial propertie s to the a c ceptability o f food dishes in which the raw mate rials are used. The function wa s included in the linear programming model together with nutritional c onstraints to p rovide a quantitative model for the selection of raw material mixe s for nutritious and a c c e ptab l e Thai food dishes.