Menu planning model for Malaysian Boarding School using self-adaptive hybrid genetic algorithm
Malnutrition problem is the gravest single threat to the world's public health today. Statistics have showed that the number of under-nourished and over-nourished children and adolescents is increasing day by day. Thus, proper menu planning process among menu planners or caterers is important t...
Disimpan dalam:
| Pengarang Utama: | |
|---|---|
| Format: | Thesis |
| Diterbitkan: |
2011
|
| Subjek-subjek: | |
| Capaian Atas Talian: | http://eprints.uthm.edu.my/3931/ http://eprints.uthm.edu.my/3931/1/Siti_Noor_Asyikin_Mohd_Razali.pdf |
| Penanda-penanda: |
Tambah Penanda
Tiada Penanda, Jadilah orang pertama menanda rekod ini!
|
| Ringkasan: | Malnutrition problem is the gravest single threat to the world's public health today.
Statistics have showed that the number of under-nourished and over-nourished children
and adolescents is increasing day by day. Thus, proper menu planning process among
menu planners or caterers is important to avoid some diet-related diseases in the future.
Manual calculation of menu planning is unable to consider macronutrients and
micronutrients simultaneously due to complexities of data and length of time. In this
study, self-adaptive hybrid genetic algorithm (SHGA) approach has been proposed to
solve the menu planning problem for Malaysian boarding school students aged 13 to 18
years old. The objectives of our menu planning model are to optimize the budget
allocation for each student, to take into consideration the caterer's ability, to fulfill the
standard recommended nutrient intake (RNI) and maximize the variety of daily meals.
New local search was adopted in this study, the insertion search with delete-and-create
(ISDC) method, which combined the insertion search (IS) and delete-and-create PC)
local search method. The implementation of IS itself could not guarantee the production
of feasible solutions as it only explores a small neighborhood area. Thus, the ISDC was
utilized to enhance the search towards a large neighborhood area and the results indicated
that the proposed algorithm is able to produce 100% feasible solutions with the best
fitness value. Besides that, implementation of self-adaptive probability for mutation has
significantly minimized computational time taken to generate the good solutions in just
few minutes. Hybridization technique of local search method and self-adaptive strategy
have improved the performance of traditional genetic algorithm through balanced
exploitation and exploration scheme. Finally, the present study has developed a menu
planning prototype for caterers to provide healthy and nutritious daily meals using simple
and friendly user interface. |
|---|