Article Preview
TopIntroduction
One problem with the shipping industry is the sudden failure of vessels at sea due to maintenance-related problems, which may result in incidents causing enormous loss of lives, limbs, valuable properties, operating hours and customer goodwill. Other consequences may include cost of ship retrieval and customer claims; rise in company insurance costs; and business survivability problems. Considering the running of a large fleet of ships, clearly, maintenance-related failures at the field of operation pose serious management problems. One solution approach to this problem is to apply a preventive maintenance system for a fleet of transportation ships. Associated with such preventive maintenance system are the following issues: (1) supply of spare parts for repairs or replacement; (2) prediction of an appropriate time to return an operating ship for preventive maintenance; and (3) provision of maintenance centres with adequate capacity to simultaneously handle arriving ships for preventive maintenance activities.
The first issue has been traditionally handled by defining and solving a spare parts inventory control problem. The second too has attracted an enormous amount of research effort. Consequently, a large number of models for predicting the period that a ship is due for preventive maintenance has been developed (Zhou et al., 2004). In a situation with small fleet of ships and large number of parallel maintenance stations, the third issue too poses no problem as every ship arriving for maintenance may immediately receive attention. In real life, this is hardly the case. Maintenance facilities for setting up even a single preventive maintenance station are very expensive. Consequently, transportation firms with many parallel maintenance stations are not common worldwide. Most reported firms have a large fleet of aircrafts or ships but a few parallel maintenance stations (Alfares, 1999; Cheung et al., 2005). In this case, many aircrafts/ships may arrive and remain in a long queue waiting to be maintained with obvious losses in revenue and heavy-operating capital tie-down.
Judicious decision has to be taken as to which of the waiting ships should first be maintained on the available stations before the rest receive the attention of the maintenance personnel. This is the preventive maintenance-scheduling decision problem for a fleet of ships. If suitable decision parameters are not selected and then an inappropriate decision model applied to derive schedules, a firm’s operations may suffer serious setback, as many high profit-yielding ships may tend to wait too long at a maintenance system. Others may fail to pick up or deliver cargo on time which may lead to heavy demurrage and contract penalty charges as well as loss in customer goodwill. It is a worldwide problem. Many nations pay close attention to this problem because of global competition. The quality of shipping services and the revenue are greatly enhanced with good preventive maintenance schedules. The maintenance scheduling methods utilised in solving problems include heuristics (Kabi et al., 2002), deterministic, multi-population cultural, multiobjective, fuzzy reliability, stochastic programming (Duffuaa & Al-Sultan, 1999; Mohanta et al., 2004; Westman et al., 2001a), memetic algorithm (Burke & Smith, 1999; Burke et al., 1998), integer programming (Lee, 1991; Dapazo & Merrill, 1975), probabilistic programming, genetic system, games theory, message-passing interface portable, tabu search (Huang et al., 2004) and genetic-evolved fuzzy approaches (Huang, 1997,1998). Billinton and Abdulwahab (2003) and others have presented probabilistic approaches to solving power generator and hydraulics problems. The use of artificial intelligence in maintenance scheduling was extensively reviewed by Dahal and McDonald (1997b). The use of simulated annealing, linear programming and hybrid evolutionary technique have also gained the interest of researchers (Wasa et al., 1999). Mathematical programming has also been used to schedule maintenance by Al-Khamis and Yellen (1995).