Journal of the Operational Research Society

Dyner, Isaac; Smith, Ricardo A.; Peña, Gloria E.

doi: 10.1057/jors.1995.165pmid: N/A

AbstractA System Dynamics model to simulate the substitution of installed household appliances by more efficient ones is presented. The model allows the construction of scenarios and also the analyses of several other issues such as: alternatives for technology penetration, electricity consumption growth, gas consumption growth and effects of pricing policies on various energy demands. The proposed methodology has been applied to assist the decision process in relation to gas penetration policies. The model also supports policy making on energy efficiency and it allows the calculation of total energy savings under different scenarios. Furthermore, government underpricing policies on tariffs and appliance acquisition (longer loan terms, lower interest rates and grace periods) may also be analysed.

Johnston, Robert Bruce

doi: 10.1057/jors.1995.166pmid: N/A

AbstractThe idea that operations in batch repetitive manufacturing can be managed using formal manipulations on representations of the manufacturing practice is at the heart of an approach called Computer-Aided Production Management. It is argued that the plausibility of this approach arises from certain widely held, but now controversial, views about the nature of purposeful human activity, which assume the need for a central representational model. Alternative theories of cognitive activity can be used to clarify the approach of a recently influential group of production management techniques, known as Lean Production, which attempt to use changes to the situation of production as their leverage point. It is claimed that implementation of these ideas depends on rendering desirable practices tacit, and this idea is illustrated using a case study.

Bohoris, G. A.; Thomas, J. M.

doi: 10.1057/jors.1995.167pmid: N/A

AbstractThis paper discusses the approach taken in solving a specific depot staffing and routeing problem. The problem differs from the normal routeing exercise in that the route passes through a number of outlets where monetary collections are made. Owing to the ever increasing risks of carrying money there is a requirement for the money to be deposited at banks along the route. Consequently, the more traditional routeing and scheduling methodologies and software are not easily applied to the problem. This paper details the approach used and highlights some of the good points and the pitfalls in the proposed solution.

Boland, N. L.; Ernst, A. T.; Goh, C. J.; Mees, A. I.

doi: 10.1057/jors.1995.168pmid: N/A

AbstractWe consider networks in which two different commodities have to be transported across undirected arcs, subject to a shared capacity on the arcs. For each arc and commodity there is an associated non-linear cost that depends on the amount of the commodity transported across the arc. The aim is to minimize the sum of the costs over all arcs and commodities. Efficient algorithms for solving this problem for two types of objective functions will be presented: in the first the cost depends on the absolute value of the flow and in the second the cost is a quadratic function of the flow. Previous work on multi-commodity flow has concentrated on linear cost problems or tackled non-linear cost problems with Lagrangian relaxation methods and other more general techniques. The algorithms in this paper, on the other hand, provide a very efficient way of dealing with two types of non-linear two-commodity optimal flow problems.

Clark, Alistair R.; Armentano, Vinicius A.

doi: 10.1057/jors.1995.169pmid: N/A

AbstractIn this paper we propose a heuristic for the resource-capacitated multi-stage lot-sizing problem with general product structures, set-up costs and resource usage, work-in-process inventory costs and lead times. To facilitate the functioning of the heuristic, we use the formulation of the problem based on Echelon Stock in a rolling horizon scheme. The heuristic first obtains a reasonable solution to the corresponding uncapacitated problem and then tries to attain capacity feasibility by shifting production backwards in time. The concept of echelon stock makes the task of checking the inventory feasibility of proposed shifts easier than would be the case with conventional installation stock. The heuristic is first tested computationally for problems with a five-component product structure over a 12 period planning horizon for which optimal solutions were available and for which optimality precision guarantees were also obtained via Lagrangian Relaxation. The heuristic's performance is also explored with two different 40-component product structures, with high and low set-up costs, and is compared with the Lagrangian precision guarantees.

Hindi, K. S.

doi: 10.1057/jors.1995.170pmid: N/A

AbstractA capacitated dynamic lot-sizing model, where the costs incurred are a start-up cost for switching the production facility on and another reservation cost for keeping the facility on, whether or not it is producing, is considered. The resulting scheduling problem is NP-hard. An efficient shortest path model of the uncapacitated version of the problem is developed. This model is then included, via a redefinition of variables, into a tight capacitated model; tight in the sense that sharp lower bounds can be produced from it. The lower bound problems are solved efficiently by recovering the shortest path structure through column generation, and effective upper bounds are generated by solving a small capacitated trans-shipment problem. The results of computational tests to verify the computational efficiency of the resulting solution scheme are presented.

Narahari, Y.; Sundarrajan, P.

doi: 10.1057/jors.1995.171pmid: N/A

AbstractFork-join queueing systems offer a natural modelling paradigm for parallel processing systems and for assembly operations in automated manufacturing. The analysis of fork-join queueing systems has been an important subject of research in recent years. Existing analysis methodologies—both exact and approximate—assume that the servers are failure-free. In this study, we consider fork-join queueing systems in the presence of server failures and compute the cumulative distribution of performability with respect to the response time of such systems. For this, we employ a computational methodology that uses a recent technique based on randomization. We compare the performability of three different fork-join queueing models proposed in the literature: the distributed model, the centralized splitting model, and the split-merge model. The numerical results show that the centralized splitting model offers the highest levels of performability, followed by the distributed splitting and split-merge models.

Hill, Roger M.

doi: 10.1057/jors.1995.172pmid: N/A

AbstractThis paper considers a product subject to a period of increasing demand, according to a general power law, followed by a period of level demand. By assuming that the Economic Batch Quantity (EBQ) is used as soon as level demand is reached, we derive the optimal production/purchasing policy for use at any stage during the period of growth. The optimal approach is compared with that of the Silver-Meal heuristic.

Afif, Mohamed; Hifi, Mhand; Paschos, Vangelis Th.; Zissimopoulos, Vassilis

doi: 10.1057/jors.1995.173pmid: N/A

AbstractWe solve approximately the minimum set covering problem by developing a new heuristic, which is essentially based on the flow algorithm originally developed by Ford and Fulkerson. We perform a comparative study of the performances (concerning solution qualities and execution times) of the flow algorithm as well as of the natural greedy heuristic for set covering originally studied by Johnson and Lovász.

Agrafiotis, G. K.; Tsoukalas, M. Z.

doi: 10.1057/jors.1995.174pmid: N/A

AbstractIn this paper a class of correlated cumulative processes, Bs(t) = ∑N(t)i=1Hs(Xi)Xi, is studied with excess level increments Xi⩾s, where {N(t), t ⩾0} is the counting process generated by the renewal sequence Tn, Tn and Xn are correlated for given n, Hs(t) is the Heaviside function and s⩾0 is a given constant. Several useful results, for the distributions of Bs(t), and that of the number of excess (non-excess) increments on (0, t) and the corresponding means, are derived. First passage time problems are also discussed and various asymptotic properties of the processes are obtained. Transform results, by applying a flexible form for the joint distribution of correlated pairs (Tn, Xn) are derived and inverted. The case of non-excess level increments, Xi < s, is also considered. Finally, applications to known stochastic shock and pro-rata warranty models are given.