Supachai Pathumnakul

An approach based on digital image analysis to estimate the live weights of pigs in farm environments

We develop a machine vision-based method to estimate the live weight of pigs.The proposed approach provides circumstance-free image processing.It is practical to apply on farms without interrupting the routine life of pigs. In this study, an estimation system for the live weights of pigs is proposed that could be practically employed in a real farm environment without disturbing the animals. This approach is based on computer-assisted visual image capture and a supervised learning algorithm known as vector-quantized temporal associative memory (VQTAM). The method is composed of three parts, which are boundary detection, feature extraction, and pattern recognition. To identify an images edge, a method that is based on user interaction via mouse-clicking on the pig image is employed to avoid edge detection errors if the pigs image and its background are not in contrast. Two image features, (1) the average distance from the pigs centroid to the boundary points and (2) the pigs perimeter length, are extracted and used as the inputs of VQTAM. Next, the solutions from VQTAM are improved by an autoregressive model (AR) and locally linear embedding (LLE). This approach has been examined using a specific farm for a case study. The results indicate that the method based on VQTAM and improved by LLE provides the most accurate prediction with an error rate of less than 3% on average.

Pig procurement plan considering pig growth and size distribution

Different pig size distributions in fattening units can prevent a pig procurement plan from achieving optimal results. Plans that fail to consider the pig size distribution and pig growth are not likely to be able to cost-effectively satisfy demand for each pig size. This paper demonstrates the use of a heuristic algorithm, pig size distribution, and pig growth to create a procurement plan. The performance of the developed procurement method is compared to the traditional practices of a company studied here. The results indicate that the company is likely to save approximately 9.52% of procurement costs by adopting the proposed method. The same problems were also investigated at an industrially relevant scale, and the computational time of the proposed heuristic was found to be reasonable. Thus, the pig industry is likely to benefit from the method developed here.

Algorithm for minimizing weighted earliness penalty in single-machine problem

In this paper, the problem of minimizing the weighted earliness penalty in a single-machine scheduling problem is addressed. For this problem, every job is assumed to be available at time zero and must be completed before or on its deadline. No tardy job is allowed. Each job has its own earliness penalty and deadline. The paper identifies several local optimality conditions for sequencing of adjacent jobs. A heuristic algorithm is developed based on these local optimality conditions. Sample problems are solved and the solutions obtained from the heuristic are compared to solutions obtained from the heuristics developed by Chand and Schneeberger. Also, comparisons are performed between the solutions obtained from the heuristic and the optimal solutions obtained from a mathematical modeling approach for problems involving 10 and 15 jobs. The results show that the developed heuristic produces solutions close to optimal in small size problems, and it also outperforms the Chand and Schneebergers method.

Harvest scheduling algorithm to equalize supplier benefits

Harvest scheduling for a group of sugar cane farmers is studied.The objective is to maximize yield with fair benefits for all farmers.Mathematical model and heuristic are developed to solve the problem.Heuristic is applied to an industrial case study in Thailand. In this study, the harvest scheduling problem of a group of cane growers in Thailand is addressed. Each member in a group is required to consistently supply sugar cane to a mill for the entire harvest season. However, the current scheduling does not account for the time-variant cane production of each cane field, which leads to unequal opportunities for growers to harvest. A portion of growers could have the opportunity to harvest in periods that provide higher sugar cane yields, while others in the same group do not. This inefficient harvest scheduling causes conflicts between growers and unnecessary loss of sugar cane and sugar yields. An artificial neural network is applied to estimate cane yield over a harvest season. Then, an optimization model and a heuristic algorithm with the objective of maximizing the estimated sugar cane yield under the condition of fair benefits for all of the growers in the group were developed to determine the most suitable sugar cane harvest schedule for a cluster of sugar cane fields. For the heuristic, the initial solution is first constructed based on the yield trends, and the solution is then improved by the tabu search approach. The results indicated that there are potential benefits from applying the model to cane scheduling within a group of heterogeneous yield trends. Sensitivity analysis showed that the more that the yield trends in a group differ from one another, the higher the benefit the group is likely to gain from adopting the proposed framework.

Determination of the locations and capacities of sugar cane loading stations in Thailand

In this paper, we address the problem of the location of sugar cane loading stations in Thailand. A loading station is a facility for collecting cane from small farmers; the cane is then transported to a sugar mill by a large truck. An improperly located loading station can result in high investment and transportation costs in the sugar industry. A mathematical model and a heuristic algorithm were developed to determine the suitable capacity of existing loading stations, the locations and capacities of new loading stations and the allocations of cane field harvests to each loading station. The model accounted for variations in the cane yield of each field during the harvesting periods and between crop years. The objective function was the minimization of the associated costs, including the investment costs, the transportation costs and the cost of the cane yield loss if the cane is not harvested at an optimal time. The performance of the developed heuristics was assessed under various scenarios. The results were shown to deviate slightly from the solution to the mathematical model. The sensitivities of the solutions under variations of the transportation cost, yield loss cost and investment costs were studied. The model was also applied to an industrial case study. A relevant and accurate solution was obtained.

Using an artificial neural network to forecast the market share of Thai rice

In this paper, the artificial neural networks (ANN) is used to estimate the market share of Thai rice in the global market. Two models are formulated under two assumptions. First, the market share depending on exporting prices of rice of Thailand, Vietnam, India, USA, Pakistan, China. Second, only the export prices of rice from Thailand, Vietnam, USA, and China are considered. The export prices are used as input parameters, while the market share of Thais rice in the global market is the only output parameter of the models. Annual data from 1980 to 2005 are gathered from United States Department of Agriculture (USDA) and Food and Agriculture Organization of the United Nations (FAO). The study showed that the second model provide more promising results with the minimum mean absolute percent error (MAPE) of 4.69% and the average MAPE of 10.92%.

Integrating a multiple crop year routing design for sugarcane harvesters to plant a new crop

Algorithms for designing sugarcane harvesters routing in multiple years are proposed.Harvesters travelling cost and sugarcane yield are considered.Algorithms are developed based on mathematical model.Fields that shared the same harvester should grow similar maturation time sugarcane. This paper discusses the integration of a multiple crop year routing design for a sugarcane harvester and planning of the planting of a new crop. A multiple crop year routing design (i.e., a three year harvesting plan) for a sugarcane crop was formulated and solved by the use of heuristics based on a VRPTW mathematical model (HVRPTW) and a dynamic programming algorithm (HDPA). The three-year harvesting period was determined from the number of years that sugarcane can normally be harvested after a crop is planted in Thailand (one planted crop and two ratoons). The model solution consisted of the harvesting sequence, the harvesters travelling routes, the harvest starting time and the number of harvesters required. The results of two methods were compared with respect to the maximum profit and computational time. The results showed that solving the problem using HDPA reduced the maximum profit by only 0.28% on average from the solution provided HVRPTW, and the average computational time was also reduced dramatically. The multiple crop year routing design was integrated with the planting of a new crop to ensure that it contained an ideal solution for the 3rd year plan so it would be effective for all three years. We recommend that the growers use a sugarcane cultivar with a similar maturation time in all of the fields that shared the same harvesters route to maintain the ideal routes. Furthermore, the same agricultural practices must be applied to all of the sugarcane crops, such as the planting method, cultivar and fertilization.

Sugarcane Harvester Planning Based on the Vehicle Routing Problem with Time Window (VRPTW) Approach

This chapter addresses the sugarcane mechanized harvesting problem. A mathematical model based on the vehicle routing problem (VRP) with time window (VRPTW) is formulated to obtain optimal harvesting plan for harvesters given specific time windows (TWs). The harvesters can operate only within the specific TW. The objective of the model is to determine the suitable sugarcane harvesting plan of a harvester. The model is applied to solve 18 generated sugarcane harvesting scenarios. Each scenario consists of 15 sugarcane fields and 20 harvesting periods. The solution comprises the harvesting sequences, traveling routes, harvesting periods, and harvest starting time. The model provides the optimal solutions which can be applied for sugarcane growers in Thailand and other similar regions.

Impacts of using relative weights in multiple criteria decision making: a comparative study between independent- and overlapping-criteria decision problems

Multiple criteria decision analysis (MCDA) methods have been widely employed in real-life decisions. An assumption generally seen is that each criterion plays a role in determining the result according to its relative weight. However, signs of imprecision in relative weights have been implied in the literature, and this may indicate that they do not always provide intuitive and satisfactory conclusions. Most MCDA methods, furthermore, were performed regardless an existence of overlap among criteria although the decision theory reports the risk of obtaining a misleading conclusion from this. Due to the lack of empirical study to demonstrate such issues, this study investigates and supports the proposition that the relative weights do not precisely reflect actual contributions of decision criteria particularly when overlaps among criteria exist. Moreover, the use of the weights in such situation, through typical additive and multiplicative aggregation methods, is likely to generate a counter-intuitive or unsatisfactory conclusion.

Using an artificial neural network and a mathematical model for sugarcane harvesting scheduling

In this paper, the sugarcane harvesting scheduling problem, in the northeast region of Thailand, is addressed. Since there are many small size farmers are participated as suppliers of the sugarcane mill, a harvesting schedule, which could provide the maximum production yield for sugarcane mill and also equal opportunity for farmers to harvest at their suitable time are required. A model, which is the combination of an artificial neural network (ANN) and a mathematical model, is proposed to solve the problem. The ANN is used to forecast sugarcane yield of each plantation over harvesting season. Then, the forecasted values are used by the mathematical model to find the optimal harvesting schedule. The objective function of the proposed mathematical model is to maximize the total sugarcane yield; meanwhile the harvesting scheduling maintains the equality among farmers in the group. The application of the model is also investigated with an example problem.