A Decision Support System for Scheduling Partial Harvesting in Aquaculture

Introduction

Aquaculture is a risky and potentially lucrative business venture. Seasoned aqua-farmers recognize that choosing an appropriate harvesting strategy is imperative to remain competitive and profitable in this business. A variety of computational models have been developed to assist aqua-farmers in determining the optimal harvest schedule in aquaculture operations (Yu & Leung, 2005; 2009). Bjorndal (1988) developed the first theoretical harvest model for aquaculture, using the optimal control framework. Arnason (1992), Heaps (1993, 1995) and Hean (1994) extended this model to take into account the effects of feeding, density-dependent growth and potential culling, and release cost, respectively. Shaftel and Wilson (1990) proposed a harvesting model using the mixed-integer linear programming approach. Forsberg (1996, 1999) developed an alternative multi-period linear programming approach. Leung and Shang (1989) employed the Markov decision approach for solving the harvesting problem in shrimp culture. Yu and Leung (2005, 2006) and Yu et al. (2006, 2009) developed an array of harvest models, primarily for shrimp aquaculture, using the network-flow modeling approach. These or similar models/modeling methods have been applied to a variety of aquacultural crops, including farmed salmon (Bjorndal, 1988), Nile tilapia (Springborn et al., 1992), giant clam (Leung et al., 1994; Hean and Cacho, 2002), carp (Talpaz and Tsur, 1982), catfish (Cacho, 1991), sea bream (Pascoe et al., 2002), crab (Figueiredo et al., 2008), prawn (Leung and Shang, 1989; Pascoe et al., 2002), shrimp (Karp et al., 1986; Hochman et al., 1990; Spaargaren, 1999; Tian et al., 2000; Pascoe et al., 2002; Yu and Leung, 2005; Pathumnakul et al., 2007; Yu et al., 2006; Kam et al., 2008) among others. Generally speaking, continuing research efforts have been made to closely mimicking the operations management of actual aquaculture enterprises. Interested readers could refer to Yu and Leung (2005, 2006) for a comprehensive review on the current status of research on optimal harvesting in aquaculture. The purpose of this chapter is to describe a decision support system for the determination of optimal partial harvesting strategy recently developed by the authors. The partial harvesting decision support system (PHDSS) is designed to assist aquaculture enterprises utilizing partial harvest (i.e., partially harvest the crop during the growout cycle) in their operations. The system automatically solves for the best partial harvesting scheme and compares it to the corresponding single-batch harvest to help aqua-farmers evaluating and implementing partial harvesting. The system is designed with great flexibility. It can be readily tailored to reflect the specific bio-economic characteristics of the aqua-farms and meet different management objectives. Single-batch harvest can be viewed as a special case of partial harvest. The PHDSS indeed could evaluate the financial implications of both single-batch harvest and partial harvest for aquaculture operations. The PHDSS is implemented in a spreadsheet form, using Microsoft Excel. It requires the use of the Solver add-in, which is available freely in Microsoft Excel.