Comparing an ant colony algorithm with a genetic algorithm for replugging tour planning of seedling transplanter

Highlights

• Optimizing the replugging tour was used to improve the efficiency of transplanter.

• An ACA algorithm and a GA were proposed for optimization of the replugging tour.

• ACA algorithm was proved to be more efficient for optimizing the replugging tour.

Abstract

The important task of replugging bad or missing cells with healthy seedlings in greenhouses is carried out by automatic transplanters. Grippers of such transplanters spend a considerable amount of time shuttling between the source and target trays during replugging. Therefore, work efficiency of transplanters can be significantly improved by tour planning. In this study, performances of the ant colony algorithm (ACA), the genetic algorithm (GA), and the common sequence method (CSM) in replugging tour planning were compared. Two types of seedling trays, with 50 and 200 cells, were used. The ACA and the GA were found to have more advantages than the CSM in total tour lengths for one tray. Moreover, the ACA performed better than the GA when the numbers of empty cells and healthy seedlings in the target and source trays, respectively, increased. When a 20 × 10 tray was used, the average length of the ACA decreased by 6000.9 mm compared with that of the GA and by 13058.4 mm compared with that of the CSM after finishing 40 empty cells in one tray. The average run times of the GA and the ACA in MATLAB (R2012a) were 0.32 and 0.94 s, respectively. These results meet real-time operation requirements.

Introduction

Manual transplantation in greenhouses is labor-intensive and costly, so growers have become increasingly interested in using automatic seedling transplanters. In 1987, researchers at Rutgers University designed a robotic transplanter that could transplant seedling plugs from high-density trays to low-density trays (Ting et al., 1992). Since this development, various end effectors have been proposed (Ting et al., 1990, Kim et al., 1995, Simonton, 1991). Several research groups have used machine vision systems to monitor seedling transplantation (Ryu et al., 2001, Beam et al., 1991, Tai et al., 1994). Commercial transplanter systems have also been developed and applied in greenhouses (Tong et al., 2013). In recent years, the commercial value and advantages of automatic transplanters in greenhouses have attracted the attention of researchers in China. Researchers in different fields (including machine vision, mechanical engineering, and control engineering) have worked together to develop efficient automatic transplanters (Jiang et al., 2009). Replugging bad or missing cells with healthy seedlings can now be efficiently done by automatic transplanters in greenhouses, and they also conduct preprocessing for most transplanting work on farmlands. The path along which the gripper moves during this process is defined as the replugging tour. Researchers have made great efforts to improve the efficiency of this process by focusing on aspects such as gripper modification (Guanjun et al., 2009, Feng et al., 2013), and optimization of the control systems (Qiao et al., 2013). The gripper spends a significant amount of time shuttling between source tray (the tray providing healthy seedlings for replugging) and target tray (the tray that needs to be emptied and replugged) while replugging. The work efficiency could be improved without increasing the cost by using a shorter replugging tour. To achieve this goal, proper planning of the replugging tour is necessary. Replugging tour planning was first proposed by our research group in 2012. There are no other published studies that deal directly with this problem. It is a combinatorial optimization problem (COP) but different from the traveling salesman problem (TSP). In the TSP, each city must be visited once and just once. Different tours consist of the same cities and follow different arrangements of all cities. TSPs have been solved by using intelligence algorithms, such as branch and bound (Shakila et al., 2013), self-organizing neural network (Masutti, 2009), artificial immune system (Qi et al., 2008), artificial bee colony algorithm (Hu, 2009), simulated annealing (Lopez-Ibanez et al., 2013), genetic algorithm (GA) (Choi et al., 2003), and ant colony algorithm (ACA) (Tsai et al., 2003, Sarhadi and Ghoseiri, 2010, Bai et al., 2013). The GA and ACA are considered in this study to resolve the replugging tour problem because of their better performance as reported in previous studies.

The GA, which is based on the principle of survival of the fittest and natural selection (Booker et al., 1989), is a promising tool for solving COPs (Morimoto et al., 1997, Scheerlinck et al., 2010, Chen, 1997). In GAs, the search space of a problem is represented as a collection of individuals. These individuals are represented by character strings, which are referred to as chromosomes. Crossover and mutation of chromosomes are the basic operations of GAs. The purpose of using a GA is to identify the best individual in the search space. ACAs emulate the behavior of real ant colonies when they search for food from their nest to food sources. Ants can find the shortest route from the food source to their ant hills without using their sense of sight (Keskinturk et al., 2012). Updating the amount of pheromones in the routes that the artificial ants pass through at the end of their tours is the basic operation of ACAs. ACAs are effective optimization methods in many applications (Tang et al., 2013, Rada-Vilela et al., 2013, Mavrovouniotis and Yang, 2013, Ghafurian and Javadian, 2011, Liu et al., 2013, Gao et al., 2013).

The present study aims to plan the replugging tour for automatic transplanters in greenhouses. The specific objectives of this study are as follows: (1) to develop algorithms for planning the replugging tour of automatic transplanters in greenhouses, (2) to select optimal values of parameters used in the algorithms, and (3) to compare the relative performances of the algorithms.