A Decision Support System for Optimizing Nutrient in Daily Lunch Intake among College Students in Cagayan de Oro City, Philippines using Linear Programming

Objectives: This study aims to develop a decision support system to ﬁnd food choices that will optimize nutrient intake among college students given their food choices and budget by formulating and solving linear programming models. Methods: Survey interviews among college students in various colleges and universities around Cagayan de Oro City were conducted to determine the foods which they usually eat as well as their usual budget for lunch. Moreover, cafeterias and fast-food chains located near school areas were also visited to ascertain the kind of foods they serve to students as well as the corresponding prices per serving of these foods. Using this information such as costs per food serving, nutrient content, student meal budget and their food preferences, linear programming models are then formulated and solved while making sure that individual nutrient requirements are met. The models are then incorporated into an excel-based program to present a user-friendly decision support system (DSS) for determining lunch menu options. Findings: Results of the study show various lunch food combinations for male and female college students. Recommended food menu costs vary according to sex, age, BMI classiﬁcation and level of daily activity. For example, the lowest for normal BMI male between 14-18 years old is PhP 40 while the lowest food menu cost for 14-18 years old female with normal BMI is PhP 45. In particular, the DSS developed in this study allow students to input their preferences such as choice of drink, viand, number of servings preferred and budget allotted for a particular meal, as well key-in their personal attributes (such as height, weight, physical activities engaged in) which are then used as basis for computing the recommended lunch menu options. Novelty. This study provides students the guide for a more nutritionally-ﬁlled lunch menu that is within their budget.


Introduction
With the transition from secondary school to college or university, as independency increases, students are constantly challenged to make healthy food selections (1) . Such transition into young adulthood is frequently a period of unhealthy lifestyle where young people could assume long-lasting health behavior habits. In particular, college students are exposed to unhealthy eating habits leading to body weight gain, and make their independent food choices, sometimes based on cost of food and availability of fast food. Diet formulation among humans generally requires, first and foremost, defining the food to be considered in the formulation and consequently, analyzing their corresponding nutrient composition (2) .
Varying age groups have different energy requirements (3) . In the Philippines, college students are usually composed of individuals whose ages are between 16-21 years old. The 8 th National Nutrition Survey conducted by Philippines' Department of Science and Technology in 2013 reports that in every 100 adolescents (aged 11-19 years old), about 17 were found to be underweight and about 5 were overweight (4) . The survey report also pointed out a significant increase in the prevalence of underweight among adolescents as compared to the 2008 Nutrition survey. Overweight adolescents, on the other hand, significantly decreased in the 2013 survey results.
The World Health Organization (WHO) Food and Nutrition Security Profiles mentioned that although the Philippines has experienced growth in per-capita gross domestic product (GDP) and dietary energy supply (DES), the dietary quality among Filipinos has remained poor and based on cereals. The poor quality of diet has contributed to high levels of stunting and underweight among young children. In addition, socioeconomic inequalities have been highly associated with malnutrition, and inadequate access to improved sanitation and high levels of food inflation have also contributed to malnutrition. Moreover, WHO reported that one-third of adults are overweight, and obesity represents an emerging issue because of unbalanced and calorie-dense diets. The reduction of physical activities among adolescents and adults have further contributed to the increase in overweight Filipinos (5) .
A healthy diet is basically one that includes the appropriate number of servings in essential food groups. College students spend a large part of their time at school, where they may consume one or more meals or snacks. An adequate diet that meets nutritional requirements is crucial for healthy physical growth and cognitive development during adolescence and can also have an influence on health and work capacity later in these students' life (5) . College or University populations are vulnerable in their eating habits for various reasons. Students might be deficient in their knowledge of healthy food selections that could negatively influence their eating habits. Financial aspects might also play a role, as fats and sweets cost less, whereas many healthy foods cost more, and increased financial concerns are associated with worse health. Planning for a nutritious meal at school may provide opportunities to students to improve their dietary intake, and in turn, help them achieve their full developmental potential.
Only 31.7 percent of Filipino households met 100 percent of the energy recommendation. The Estimated Average Requirements (EAR) for niacin and protein was met by 86.4 and 62.7 percent of households, respectively. The proportion of households meeting the EAR for the rest of the nutrient ranged from 25 to 35 percent except calcium (15.2%) and iron (8.8%). In rural areas, the proportion of households meeting the energy intake was slightly higher than in urban areas. However, the proportion of households meeting the EAR for all nutrients was higher in urban areas, except for calcium and vitamin C. Among regions, the proportion of households meeting 100 percent of the recommended energy intake was highest in CAR (40.8%) and lowest in SOCCSKSARGEN (25.8%). Regions with the highest proportion of households meeting the EAR for nutrients were NCR for protein (72.4%), CAR for iron (19.4%), vitamin A (35.4%), thiamin (49%) and riboflavin (34.8%) and Cagayan Valley for calcium (26.5%) and vitamin C (44.5%) (6) .
The main goal of nutrition plans is to obtain the appropriate and necessary nutrition to remain healthy, to be physically prepared and to lead a healthy life. For this reason, to promote the health level of a society, and the attitudes of its people, must be taken into account. Given that one of the main goals of colleges and universities is to broaden the knowledge of the society, the enhancement of the nutrition attitudes, knowledge and practices of the students is of high importance, as this will subsequently lead to a more food-conscious society and healthier people. Some studies have shown that most students are not familiar with the healthy foods needed for their body in different conditions (1,4) . Research showed that the majority of students (83.6%) eat three meals during the day regularly and no difference was found between men and women (6) .
According to the 8 th National Nutrition Survey Report (2) , energy inadequacy is greatly observed among adolescents. Based on their study, over a period of 20 years, the mean daily energy intake increase can only be observed among preschool children. Tanchoco (4) reviewed the formulation of positive, practical and culturally sensitive food-based dietary guidelines (FBDGs) to help Filipinos choose an adequate diet and foster wholesome food and nutrition practices to promote good health. She pointed out that there is a need to come up with a comprehensive plan that includes implementation, assessment, monitoring and reformulation of dietary patterns for Filipinos. The paper provided guidelines on how food for Filipinos must be prepared in order to meet the required nutrients.
https://www.indjst.org/ Angeles-Agdeppa, et al. (1) investigated the dietary intakes and acceptance of nutritionally balanced school meals (or nutrimeals) as compared with regular (or baseline) school meals among high school students. They found out that nutritionally balanced nutri-meals may be a healthier and acceptable alternative to regular Filipino school meals. The focus of the study was on assessing the acceptability of new school meals among Filipino students. Results of the study may be used as basis for improving dietary intakes among students.
Magbuhat, et al. (5) examined differences in food preferences and dietary intake among male and female Filipino adolescents of different nutritional status as measure by body mass index (BMI). Results of their study suggest that food preferences should be considered in the nutritional care management of malnourished adolescents. It should be noted that this study only focused on the food preferences of Filipino adolescents, considering both its effect on the nutrient intake and its effect on the nutritional status of each respondent.
Linear programming has seen real-world applications in health (7) also been applied to providing diet recommendations to human. Namoco (8) , Agdeppa-Namoco and Gican (9) and Namoco, et. al (10) conducted studies on the use of linear programming for animal feed formulation. The study presented various possible animal feed rations using ingredients which are locally available. Another interesting aspect of the study is that they used ingredients which are not necessarily commercially available, and hence do not have direct cost. However, the focus of the study is to find feeds at a minimum cost so that the authors made use of the labor cost for the feed cost. Saxena and Parasher (11) , on the other hand, developed mathematical models for animal diet formulation using artificial neural network (ANN).
Alaini, et. al (12) used linear programming method to develop low-cost cancer prevention food plans based on existing cancer dietary guidelines in order to help cancer patients achieve healthy eating style. Fletcher, et al. (13) provided a mathematical model on the relationship between acceptable diets and satisfactory nutrition. The study emphasized that the classical minimum-cost diet problem is not generally suited to most purposes in human dietetics since the objective of providing least-cost diets may lead to unpalatable diets. Linear integer programming was also used to solve a diet problem for McDonald's set menu in the work of Mohamed, et. al (14) .
There have been studies on human diet formulation using mathematical programming techniques. Similar studies on diet formulation have also been conducted for animal feeds. While mathematical programming techniques help obtain optimal diet formulation either for humans or animals, it can be seen from these existing studies that they normally deal with fairly databases of food choices and their corresponding nutritional analysis. Consequently, users may find solving these mathematical models complex and tedious.
With the advent of computer technology, computer-based solvers have been made available. These computer-based systems help make the decision-making process easier and convenient. A decision support system (DSS) provides an alternative way to reduce uncertainties that may result when solving large optimization problems. A DSS normally consists of databases, modelling functionality and a user-friendly interface.
In this study, a DSS is developed to provide for daily diet planning to optimize nutrients in daily lunch intake among college students. The study is aimed to aid students in choosing lunch that are more nutritious yet are within the current budget normally spent by these students. The study shall be limited to the food available within the school area. Furthermore, an excel-based DSS is developed in this study to provide students with a user-friendly platform for deciding which food choices for lunch to take.

Methodology
To achieve the objectives of this study, surveys and interviews were conducted among college students in various colleges and universities around Cagayan de Oro City, Philippines to determine the food that these students usually consume for lunch, including the serving sizes as well as the budget that they usually allot for their lunch. Actual visits to four universities in Cagayan de Oro, namely, the University of Science and Technology of Southern Philippines (a state university) and Xavier University, Liceo de Cagayan University and Capitol Universities (which are private universities) were made to conduct the survey. Randomly selected college students, with ages between 18-30, from these universities were asked to provide their usual food choices when eating lunch using the survey questionnaire developed and prepared by the researchers. The questionnaire also asked the student-respondents to specify their usual food budget, the activities that they are usually engaged in every day and other personal information.
Eateries and fast-food chains around these universities were also surveyed to ascertain the kind of foods they serve to students (including the ingredients used in these foods and the costs per servings). All eateries situated within a 100-meter radius from the universities were included in this study.
It is interesting to note that the food choices mentioned by students in the survey are the same food that are being served in the eateries and fast-food restaurants around the universities. Correspondingly, all these food choices are included in the https://www.indjst.org/ database designed for the decision support system developed in this study.
After all these foods have been determined, the nutrient contents of these foods as well the corresponding nutrient requirements of males and females (of varying ages) were determined through the help of published literature as well a nutritiondietetics specialist. Table 1 shows the usual food that college students eat during lunch. On the other hand, Table 2 shows the food available from fast-food chains as well as their nutrient contents and price per serving. The corresponding nutrient content and cost per serving are also shown in these tables. The nutrient analyses are based from the Philippine food composition tables 1997 (15) .   Commonly eaten lunch by college students available from fast-food chains.
It should be noted that as of this writing, 1 USD ($) is approximately equal to 50 Philippine Peso (PhP). Tables 3 and 4 show the needed energy and nutrient intake of male and female college students for different age brackets (16) . The body mass index (BMI) categories for both male and female of varying ages are also shown in the tables. The values reflected are for lunch, which is computed as one-thirds of the daily requirement for each individual.
Using the information from Tables 1 and 4, a linear programming (LP) model that will determine the right food combination which meets the required nutrient intake for college students based on their preferred food and allotted budget is then formulated. The main objective of the linear programming model is to maximize the buying power of each student's allotted budget for lunch given their preferred food choices and specified daily activities. Constraints of this LP model include the daily nutrient requirements among college students based on age group and the amount of money they allocate for a meal (daily budget for lunch).
1, i f f ood choice i ∈ I is chosen to be part o f a meal 0 otherwise where I is the set of all available food choices from Table 1 and Table 2. Then, for each j=1, 2, …, 12 (corresponding to the different age intervals for males and females), we have the following LP model.

is the budget alloted f or a meal) x i is Binary
where A U , A L , B U , B L , D U , D L , F U , F L correspond to the maximum and minimum energy, carbohydrates, protein and fat requirements for males and females of the different age intervals and varying BMI descriptions, as shown in Tables 3 and 4. To better put the results of this study into good use, a decision support system (DSS) is then designed and developed to provide an easier lunch planning guide to college students. https://www.indjst.org/

Results and Discussion
The DSS developed in this study provides suggestions on what a student may eat at lunch given his/her food preference and budget. This is attained by solving the LP model mentioned above. Table 5 shows a sample lunch menu for a male, between 14-18 years old and with normal BMI. Based on our computations, it was found out that the smallest possible cost to satisfy the minimum requirements for energy, carbohydrates,protein, and fat, is PhP 40, for food bought at the cafeteria. This food menu includes rice, chopsuey and bihon. It can be observed that as the budget is increased, the food included in the menu also changed, and correspondingly, the nutrient contents of these food choices vary. It is also interesting to note that as the budget increases, the number of food choices included in the menu also increase. For example, at PhP 40 only two (2) viands are included while at PhP 80, the menu has three (3) viands. Moreover, since all-meat viands are more expensive that those with vegetables, it can also be observed that the cheaper recommended menus are mostly vegetable-based while the more expensive ones have a combination of all-meat and vegetable viands. Additionally, rice has always been a part of the recommended food menu. Similar observations can also be made from Tables 6-8 for the suggested food compositions for both 14-18 years old male and female who are overweight.
The main goal of this study is to develop an automated system that will help college students decide which food to eat for lunch given their food preferences and budget. Hence, we have prepared a decision support system that captures the linear programming model as presented in the previous section and demonstrated by the sample suggested menus in Tables 5, 6, 7 and 8.  The following figures provide information on how the DSS can be used. Figure 1 shows the landing page of the DSS developed in this study. Here, users are required to input their details. This information is then processed and results are shown in Figure 2. The system is developed in a very user-friendly manner. Figure 3 shows how the system informs the user that it is ready to release results based on the information given by the user. On the other hand, the system prompts a warning message when the information supplied by the user is incomplete. This is shown in Figure 4. Once complete information is entered in the system, results are then displayed based on the supplied preferences. As can be seen in the following figures, various considerations were included in the setup. These include meal recommendations for male and female of various ages and activities. Similarly, the mathematical model used in the DSS developed in this study also considers different budget constraints. Sample outputs from the DSS developed in this study are shown inFigures 5, 6, 7, 8 and 9. It is worth noting that the DSS developed in this study allows users to specify their food preferences while considering their body figure and budget.

Conclusion
The linear programming models formulated and solved which were then incorporated in the DSS developed in this study provide college students a better way to choose the food they want to consume for lunch by ensuring that the required nutrients are satisfied while keeping the cost at a minimum. Moreover, students are given better insights on how these foods contribute to their wellness. The results of the study further show that these food choices vary according to their daily activities and meal budget. The introduction of the DSS enables users to have a more dynamic way of customizing their food menu, taking into consideration their food choices, individual characteristics and other personal preferences and circumstances. In particular, food costs vary according to age, sex, body figure and kind of activity of the individuals. We have seen that the lowest cost for food menu for 14-18 years old male with normal BMI is PhP 40 while for female, the cost is P45.
While results of the study provide a better way to plan meals among college students, it would be more interesting to come up with a more comprehensive system for meal planning to include people of different age brackets.