C.K.Y. Lin

Multi-objective metaheuristics for a location-routing problem with multiple use of vehicles on real data and simulated data

We address an integrated logistic system where decisions on location of depot, vehicle routing and assignment of routes to vehicles are considered simultaneously. Total cost and workload balance are common criteria influencing decision-making. Literature on location-routing problems addressed the location and vehicle routing decisions with a common assumption of assigning one route to one vehicle. However, the cost of acquiring vehicles (and crew) is often more significant than the routing cost. This notion of assigning several routes to a vehicle during the routing procedure is explored in our integrated model. We apply metaheuristics of tabu search and simulated annealing on real data and simulated data, to compare their performances under two versions: simultaneous or sequential routes assignment to vehicles. A new statistical procedure is proposed to compare two algorithms on the strength of their multi-objective solutions. Results show that the simultaneous versions have advantage over the sequential versions in problems where routes are capacity-constrained, but not in the time dimension. The simultaneous versions are also more effective in generating non-dominated solutions than the sequential versions.

A location-routing-loading problem for bill delivery services

A telecommunication service company in Hong Kong considers printing and delivering its monthly bills to well-structured and densely populated housing estates by the in-house delivery team. There is a large customer base that prefers receiving paper bills for personal or company accounting purposes. The use of in-house delivery team would be more cost-effective than the existing practice of employing the Post Office service. With the anticipated increase of housing estates in the newly developed areas, management is considering to relocate the existing bill delivery office and set up some delivery depot sites at the existing company buildings. Metaheuristic approach based on threshold accepting and simulated annealing is developed to assist in making facility location, vehicle routing and loading decisions. The computerized algorithm has been tested by the operation team and is well perceived by the management. It shows significant improvement over the existing manual approach and will be a tool useful for planning future office locations in Hong Kong.

Development of a workforce management system for a customer hotline service

Abstract An integrated workforce management system was developed for a 24-h hotline service. It first applies either a regression model or simulation model to relate the target abandonment rates with the required hourly staffing level, which is then converted into the daily requirements of senior and junior officers by use of a mixed integer programming approach. The last phase involves extending Burns and Carter algorithm to schedule the monthly roster with emphasis on both employee equity and management specifications. Implementation is under Microsoft Office 95 on a 586 PC. Comparison with the manual schedule shows significant savings in time and a higher degree of constraint satisfaction. Scope and purpose With the booming mobile telephone business, a 24-h customer hotline service in a mobile communications company faced increasing challenges in scheduling its growing workforce to be distributed into eight daily shifts. The complex constraints and goals set by the company demand a tailor-made workforce management system where existing packages cannot be directly applied. This paper reports on the development of a decision support system for the forecasts of hourly demand calls, leading to the estimation of staffing requirements and scheduling of the monthly roster. Fast rescheduling can be achieved and more scheduling constraints are satisifed. Useful management information is generated to enable better understanding of the manpower supply and demand situations. The computerization process brings together the management and interdisciplinary teams to realize the tradeoffs between various scheduling constraints, and possible modifications in anticpation of future changes.

A cooperative strategy for a vehicle routing problem with pickup and delivery time windows

A vehicle routing problem with pickup and delivery time windows is studied with the objective of determining resource requirements and daily routing by minimizing the sum of vehicle fixed costs and travelling costs. The multiple use of vehicles can reduce costs. Such a cooperative strategy, which has received little attention in past literature but which occurs in practice, may generate further savings. The strategy is studied here, with the single or multiple use of vehicles. Vehicles are allowed to travel to transfer items to another vehicle returning to the depot, provided no time window constraint is violated. This is modelled by an exact integer programming formulation which includes the solutions of the independent strategy. The proposed models are compared with a construction heuristic [Lu, Q., & Dessouky, M. M. (2006). A new insertion-based construction heuristic for solving pickup and delivery problem with time windows. European Journal of Operational Research, 175 (2) 672-687] which was applied to this problem. Experiments with instances generated from real-life data and simulated data show (i) significant savings over the construction heuristic as problem size grows; and (ii) multiple use of vehicles with a cooperative strategy may achieve cost savings over the independent strategy.

A comparative study of both standard and adaptive versions of threshold accepting and simulated annealing algorithms in three scheduling problems

Abstract Local search techniques are reported to be effective for many optimization problems with arbitrary objective functions and constraints. This paper compares simulated annealing with a less frequently mentioned approach, threshold accepting. Three scheduling problems of jobs with arbitrary time lags in a single-server system are considered. The optimization criteria used are maximum completion time and mean completion time. Two adaptive versions of both techniques: the adaptive neighbourhood search and the adaptive temperature/threshold values are proposed which guide the future search according to the recent search performance. These adaptive concepts prove to be effective and may be applied into other local search techniques. Computational results show that threshold accepting algorithms achieve better or comparable performance with simulated annealing in the standard and adaptive versions. Incorporating descent approaches into threshold accepting may lead to substantial improvement.

A vehicle routing problem with pickup and delivery time windows, and coordination of transportable resources

This paper addresses a pickup and delivery problem with time window constraints, using two types of delivery resources, which allows for coordination. A heavy resource (in this case a van) may carry both delivery items and one or more units of the lighter resource (foot couriers) on its single- or multi-route assignment. Foot couriers can pick up and deliver items independently, or travel with a van on its outbound and/or return leg. Coordination between resources can save time and cut costs. The optimization of this type of coordinated logistics system has received scant coverage in the literature. The objective criterion is the total cost (i.e., the sum of fixed cost and travelling cost) of both heavy and light resources. The problem is first formulated as a mixed integer program, assuming a reasonable coordination strategy. A two-stage heuristic is designed by employing both exact approaches and an ejection chain based on modifying the shortest path algorithm implemented in a dynamic network. The two-stage heuristic is compared with optimization models for an independent strategy and a partial coordination strategy on both local and simulated instances of up to 200 customers. Under certain conditions, the benefit of coordination is realized over the independent strategy in terms of lower average total cost and usage of heavy resource.

Hospital lift system simulator: A performance evaluator-predictor

Abstract Hospitals often experience lift congestion as a result of their heavy traffic, complex user types, and relatively slow-moving lifts (due to concerns over safety). Given the increasing number of current and new hospital building blocks that consist of many storeys, a visual simulation-based decision support system (DSS) is recommended. We present the modelling approach and development of a tool capable of being used for lift performance evaluation/prediction of existing/new hospital designs. These are also applicable to other general-purpose lift systems. A new data modelling approach, based on collected empirical traffic data, was developed to estimate the inter-floor passenger traffic. The DSS is flexible enough to allow the input of any zoning policy. The integrated zoning analysis offered here has not been found in existing lift simulators. This paper is the first to model a special feature designed to disable certain lift buttons in order to ensure fair use of the lift service. We carried out field studies of two existing hospitals, and we projected lift demand for a new hospital under construction. Performances at all three hospitals with different design structures under different operational control policies and lift features are given.

An adaptive scheduling heuristic with memory for the block appointment system of an outpatient specialty clinic

This work analysed the appointment system of outpatient clinics serving multiple patient classes with different flow sequences through the multi-phase-multi-server service system. Scarce resources are doctors, nurses and medical professionals with different start times and availability. Block appointment systems are typically used in public hospitals to help regulate patient flow while minimising patient waiting time, staff overtime and waiting room congestion. The patient scheduling problem in this complex environment is formulated by a mixed integer programme (MIP). Making use of waiting time information, an adaptive scheduling heuristic is designed to improve an initial schedule iteratively by identifying procedures with large average waiting times and reassigning their related patient classes to less congested time blocks probabilistically. An impact index based on the weighted multi-objective function is developed to allow servers select an available patient for the next treatment. A memory of distinct solutions is maintained to avoid recycling. Experiments are conducted based on a case study of an eye clinic in a public hospital. Performance is evaluated by comparing with the MIP and well-known dispatching rules for job shop scheduling problems. Sensitivity analysis is conducted for increase in appointment quota, two alternative staffing plans and changes in patient class distribution.

Microcomputer-based workforce scheduling for hospital porters.

This paper focuses on labour scheduling for hospital porters who are the major workforce providing routine cleansing of wards, transportation and messenger services. Generating an equitable monthly roster for porters while meeting the daily minimum demand is a tedious task scheduled manually by a supervisor. In considering a variety of constraints and goals, a manual schedule was usually produced in seven to ten days. To be in line with the strategic goal of scientific management of an acute care regional hospital in Hong Kong, a microcomputer-based algorithm was developed to schedule the monthly roster. The algorithm, coded in Digital Visual Fortran 5.0 Professional, could generate a monthly roster in seconds. Implementation has been carried out since September 1998 and the results proved to be useful to hospital administrators and porters. This paper discusses both the technical and human issues involved during the computerization process.

Heuristic Approaches for a Scheduling Problem in the Plastic Molding Department of an Audio Company

A production scheduling problem for making plastic molds of hi-fi models is considered. The objective is to minimize the total machine makespan in the presence of due dates, variable lot size, multiple machine types, sequence dependent, machine dependent setup times, and inventory limits. Goal programming and load balancing are applied to select the set of machine types and assign mold types to machines, resulting in a set of single-machine scheduling problems. A mixed-integer program (MIP) is formulated for the general problem but could solve only small instances. A single-machine scheduling heuristic is designed to adopt a production sequence from a travelling salesman solution. The start time of every cycle is determined by a simplified MIP. Production cycles are defined to equalize the stockout times of mold types. A post-processing step reduces the number of setups in the last cycle. Results using real-life data are promising. Characteristics giving rise to high machine utilization are discussed.