文档下载 文档下载导航
设为首页 | 加入收藏
搜索 请输入内容:  
 导航当前位置: 文档下载 > 所有分类 > 自然科学 > 数学 > Linear Programming and Mixed Integer Programming in Management of Seedling Transportation
侵权投诉

Linear Programming and Mixed Integer Programming in Management of Seedling Transportation

Linear Programming and Mixed Integer

Programming in Management of

Seedling Transportation

Juho Rantala

Finnish Forest Research Institute

Suonenjoki, Finland

ABSTRACT

In this paper, the applicability of linear programming(LP) in management of seedling transportation wascompared to that of mixed integer programming (MIP). Inthe LP model, presented in an earlier paper, a linearobjective function was used as a surrogate for the actualobjective function, which is intrinsically nonlinear. In theLP model, transportation costs were determined perseedling, whereas in the MIP model they were based onvehicle loads. When the number of transported seedlingswithin a certain period decreased, for instance, due toplanting through the growth period, the computationalaccuracy of the LP model was clearly lower than that ofthe MIP model. Despite that, differences in allocation oforders between these two models were small. Thus, inthe actual business situation of Finnish nurserycompanies, standard LP seems to be an adequate tool formanagement of seedling transportation. From thestandpoint of cost-efficient seedling business, plantingthrough the growth period increased optimaltransportation costs markedly. In addition to the seedlingbusiness, these results can be utilized in other types ofbusiness dealing with analogous transportationproblems.

Keywords:nursery, transportation optimization, mixed

integer programming, linear programming,Finland.

INTRODUCTION

In Finland the forest nursery business has undergonehuge changes during the last decade, and nurserycompanies are presently getting used to a new businesssituation. Companies have been privatized, domesticcompetition has increased and the seedling business hasbecome more international. In addition, changes inlegislation concerning the seedling trade have changedthe role of middlemen, for instance, forest owners’associations (FOAs), making them more responsible tothe end-users for the quality of seedlings. Therefore, theThe author is a Researcher at the Suonenjoki ResearchStation, Finnish Forest Research Institute.

International Journal of Forest Engineering 41

middlemen are at present more clearly the real customersof nursery companies. Therefore, in this study, the word“customer” refers to middlemen. At the same time, thenumber of nursery units has decreased, which hasemphasized the importance of transportation as a part ofthe seedling business. In the future, general use ofplanting through the growth period (PTGP) might changecustomers’ requirements, in particular, for seedlingdelivery activities. One possibility for nursery companiesto respond to customers’ requirements, and at the sametime operate more cost-effectively, is to develop moreadvanced systems of transportation management.According to recent studies, costs for seedlingtransportation could be lowered markedly by utilizationof mathematical programming [10]. To study the effectsof centralized transportation planning on transportationcosts, Rantala et al. [10] applied standard linearprogramming (LP). In the LP model, transportation costswere calculated per seedling and based on the assumptionof full vehicle loads. The advantages of LP are, forinstance, short and fairly predictable solution times. LPas a method has proved to be practical especially as atool for strategic and tactical level planning oftransportation [9]. Nevertheless, for the problem examinedin the earlier paper by Rantala et al. [10], it is likely that theaccuracy of LP will deteriorate when the total number oftransported seedlings is small compared to the capacitiesof the transportation vehicles. This is due to the fact thata linear objective function was used in the formulation asa surrogate for the actual objective function, which isintrinsically nonlinear because it involves both fixed andvariable costs. When the applicability of differentoptimization methods to the optimization of seedlingtransportation was prefigured, a few aspects wereconsidered, in particular, how long the transportationdistances will be, what the transportation costs offractional vehicle loads are compared to those of full loadsand what are the sizes of seedling orders in proportion tothe capacities of transportation vehicles. Here, to evaluatethe applicability of LP by constructing an actual modelfor management of seedling transportation, integerprogramming (IP) is introduced. In the field of foresttechnology, IP has previously been used, for instance, inmodeling the optimal use of log-stacking lift trucks atwood terminals [4]. Mikkonen [8] introduced mixed integerprogramming (MIP), an adaptation of IP, as a tool forchoosing harvesting systems.

Mathematical optimization problems, which includeinteger restrictions, are usually difficult to solve andrequire a huge computing capacity and much time. WithIP models, formulation is crucial in determining whetherthe problem is solvable, and what the solution time willbe[6, 11]. Other important factors affecting solvability of

第1页

你可能喜欢

热门文档

相关文档

站点地图 | 文档上传 | 侵权投诉 | 手机版
本站所有资源均来自互联网,本站只负责收集和整理,均不承担任何法律责任,如有侵权等其它行为请联系我们.
文档下载 Copyright 2013 doc.guandang.net All Rights Reserved.  -  email