A stochastic detailed scheduling model for periodic maintenance of military rotorcraft

Purpose The purpose of this paper is to develop a stochastic detailed schedule for a preventive/scheduled/periodic maintenance program of a military aircraft, specifically a rotorcraft or helicopter. Design/methodology/approach The new model, entitled the military “periodic aviation maintenance stochastic schedule” (PAM-SS), develops a stochastic detailed schedule for a PUMA SA 330SM helicopter for the 50-h periodic inspection, using cyclic operation network (CYCLONE) and Monte Carlo simulation (MCS) techniques. The PAM-SS model identifies the different periodic inspection tasks of the maintenance schedule, allocates the resources required for each task, evaluates a stochastic duration of each inspection task, evaluates the probability of occurrence for each breakdown or repair, develops the CYCLONE model of the stochastic schedule and simulates the model using MCS. Findings The 50-h maintenance stochastic duration follows a normal probability distribution and has a mean value of 323 min and a standard deviation of 23.7 min. Also, the stochastic maintenance schedule lies between 299 and 306 min for a 99 per cent confidence level. Furthermore, except the pilot and the electrical team (approximately 90 per cent idle), all other teams are around 40 per cent idle. A sensitivity analysis is also performed and yielded that the PAM-SS model is not sensitive to the number of technicians in each team; however, it is highly sensitive to the probability of occurrence of the breakdowns/repairs. Practical implications The PAM-SS model is specifically developed for military rotorcrafts, to manage the different resources involved in the detailed planning and scheduling of the periodic/scheduled maintenance, mainly the 50-h inspection. It evaluates the resources utilization (idleness and queue), the stochastic maintenance duration and identifies backlogs and bottlenecks. Originality/value The PAM-SS tackles military aircraft planning and scheduling in a stochastic methodology, considering uncertainties in all inspection task durations and breakdown or repair durations. The PAM-SS, although developed for rotorcrafts can be further developed for any other type of military aircraft or any other scheduled maintenance program interval.

Least Expected Time Paths in Stochastic Schedule-Based Transit Networks

We consider the problem of determining a least expected time (LET) path that minimizes the number of transfers and the expected total travel time in a stochastic schedule-based transit network. A time-dependent model is proposed to represent the stochastic transit network where vehicle arrival times are fully stochastically correlated. An exact label-correcting algorithm is developed, based on a proposed dominance condition by which Bellman’s principle of optimality is valid. Experimental results, which are conducted on the Ho Chi Minh City bus network, show that the running time of the proposed algorithm is suitable for real-time operation, and the resulting LET paths are robust against uncertainty, such as unknown traffic scenarios.

A Simulation Model for LTL Trucking Network

Service network design (SND) is a part of tactical planning activities of transportation companies. Less-than-truckload (LTL) trucking industry has been steadily expanding the market share in the past decades, due to its operational flexibility and high efficiency. In order to provide flexible and robust service schedule for LTL carriers, stochasticity is explicitly taken into account when formulating the SND problem. Service schedules derived from the stochastic model show structural difference with its deterministic counterparts. This research project develops a simulation model of an LTL network, in order to evaluate the system performance of LTL network with the stochastic schedule. A set of experiments shows that the stochastic solution performs very well when it is confronted with random customer demands. Furthermore, the stochastic schedule is much better than the deterministic one in terms of the proportion of undelivered commodities.

Great Tits and Giving-Up Times: Decision Rules for Leaving Patches

In this paper I examine the hypothesis that animals employ simple decision rules in exploiting patchily distributed food. This idea arises from the supposition that animals probably cannot compute the optimal patch residence time in the same manner that is calculated in optimal foraging models. Instead they may only approximate the optimal solution using a simple, robust rule-of-thumb. I considered three rules-of-thumb that great tits foraging in a simple experimental habitat may use: a number expectation, a time expectation, and a giving-up time. The experimental habitat consisted of an operant patch in which great tits had to search for food by hopping on a perch. The probability of reward declined for successive patch hops according to a predetermined stochastic schedule. In order to maximize food intake, the great tits had to occasionally leave the patch and fly across the experimental room to reset the reward schedule by hopping on a second perch. Three different lines of investigation were followed to test which of the three departure rules were employed: (a) a statistical analysis showed that there was a strong tendency for great tits to leave the patch only after several successive hops had gone unrewarded; (b) a computer simulation showed that a simple giving-up rule could produce a distribution of patch residence times similar to that observed; (c) an experiment in which the reward schedule was manipulated, successfully altered the patch residence times in accord with predictions made on the basis of a giving-up time rule. Thus, all three tests produced evidence in favour of the giving-up time rule. Although the great tits used a giving-up time rule, a residence time would have resulted in a higher rate of intake. One potential explanation of this apparent error is that the natural food of great tits has a clumped distribution, to which a giving-up time is better suited than a residence time rule. Finally, I point out that the observed giving-up time was much more variable than might be expected if it were adjusted solely in response to the habitat rate of intake. I suggest some hypotheses to explain this large variation.