Harvey H. Millar

Using formal MS/OR modeling to support disaster recovery planning

Abstract All organizations are susceptible to a non-zero risk of experiencing out-of-course events, whether natural or man-made, that can lead to internal “disasters” with respect to business operations. Different types of events (e.g. flood, earthquake, fire, theft, computer failure) have implications for the operations of modern organizations. Hence, there is a critical need for planning and recovery strategies for the effects of disasters. Disaster recovery plans (DRPs) aim at ensuring that organizations can function effectively during and following the occurrence of a disaster. As such, they possess cost, performance, reliability, and complexity characteristics that make their development and selection non-trivial. To date, there has been little modeling of disaster recovery issues in the MS/OR literature. We believe that many of the issues involved can benefit from the application of quantitative decision-making techniques. Consequently, in this paper our contribution is prescriptive rather than descriptive in nature and we propose the use of mathematical modeling as a decision support tool for successful development of a DRP. In arriving at a final DRP, decision-makers must consider a number of options or subplans and select a subset of these subplans for inclusion in the final plan. We present a mathematical programming model which helps the decision maker to select among competing subplans, a subset of subplans which maximizes the “value” of the recovery capability of a recovery strategy. We use hypothetical situations to illustrate how this technique can be used to support the planning process.

Lagrangian heuristics for the capacitated multi-item lot-sizing problem with backordering

Abstract This paper presents two algorithms for solving a network-based formulation of the capacitated multi-item lot-sizing problem with backordering. We employ Lagrangian decomposition and Lagrangian relaxation techniques which exploit the underlying network structure of the problem. In both approaches, we exploit a transportation subproblem which guarantees a primal feasible solution at every iteration of the procedure. Further, we use a primal partitioning scheme to produce additional primal feasible solutions. Valid lower bounds are obtained at each iteration of the algorithms, thereby providing a readily available ex post measure of the quality of the primal solutions. Computational analysis shows that both algorithms are quite effective, particularly when item setup and unit backorder costs are high. We also provide a means of evaluating the potential impact of permitting backorders under various problem characteristics.

An application of Lagrangean decomposition to the capacitated multi-item lot sizing problem

Abstract In this paper we present a Lagrangean decomposition technique for solving the capacitated multi-item lot sizing problem (CMLSP). The approach decomposes the problem into a transportation problem and N independent single-item uncapacitated lot sizing problems. The algorithm applies subgradient optimization to update the Lagrangean multipliers. Each iteration of the Lagrangean procedure generates two primal feasible solutions: one from the transportation subproblem, and the other from the network flow problem resulting from a primal partition produced by the solution to the N single-item problems. The Lagrangean algorithm proves to be quite effective and stable in producing good primal and dual solutions to the CMLSP.

Planning harvesting and marketing activities for integrated fishing firms under an enterprise allocation scheme

Abstract This paper focuses on the tactical planning problem for integrated fishing enterprises operating in Canadas Atlantic Groundfish industry. A firm with a fleet of trawlers, a number of processing plants, quasi-property rights to fish in the sea, and market requirements, must coordinate harvesting and marketing strategies which will allow it to maximize potential revenue. To achieve this, we outline a large-scale linear programming model which maximizes net revenue from product sales less fleet operating costs, subject to marketing and fleet operating constraints. The output of the multi-period linear program suggest what products and their respective volumes should be marketed, and how raw fish should be caught in order to satisfy marketing requirements. We demonstrate the planning and diagnostic potential of the linear program by solving an example problem based on data obtained from a large Atlantic seafood company.

Dispatching a fishing trawler fleet in the Canadian Atlantic groundfish industry

Abstract An integrated fish-processing firm in the Canadian Atlantic demersal fishery usually owns a fleet of fishing trawlers. For a given planning horizon, the firm must find a minimum-cost fleet dispatching plan in order to satisfy demands for various species at its processing plants. In this paper, we formulate two cases of this trawler dispatch problem as mixed-integer programming models. In addition, we develop heuristics for solving both problems. Results for several sample problems, show quite favorable performance by the heuristic methods. Solution quality averaged within 2% of Lagrangean lower bounds. Also, using three test problems extracted from a firms historical fishing records, we produce in a deterministic setting, solutions which represent up to 30% improvement over the firms real-time solutions.

The impact of rolling horizon planning on the cost of industrial fishing activity

Abstract Industrial fish-processing firms must operate against a backdrop of uncertainty in the demand for fish (the raw material) and uncertainty in the demand for final products. Integrated firms must dispatch a harvesting fleet to land required catch at one or more processing plants. Typically fishing trips for individual vessels last for up to two weeks. In planning harvesting schedules, the decision maker must take into account the impact of the uncertainties mentioned above. The purpose of this research is to gain insight into the effectiveness of rolling horizon planning as a mechanism for coping with uncertainty, and to establish the nature of the relationship between the length of the frozen and planning horizons and the expected annual costs of fishing plans. Such insight will be extremely valuable to fleet harvest planners who must attempt to minimize the landed cost of raw fish. Vertically integrated fish-processing firms must dispatch their fishing fleet to meet short-term or long-term requirements for raw fish at the various processing plants. Fleet harvesting activity must be planned against a backdrop of uncertain catch rates, fish quality, and demand. In this paper, we investigate the potential for using rolling horizon planning as a strategy for managing the impact of these uncertainties, particularly in catch rates. We study the impact of the length of the planning and frozen horizons on the cost performance of fishing fleet schedules. Results for several scenarios involving variable catch rates, variable demands, and alternative rolling/planning horizon combinations are presented.

A simulation model for assessing fishing fleet performance under uncertainty

The authors present a simulation model for assessing the impact of myopia and catch rate variability on the cost performance of a coordinated fishing fleet. The trawler dispatching problem faced by integrated fish-processing firms in the Canadian Atlantic groundfish industry is modeled using the simulation language SLAM II. The results show that myopic dispatching of the fleet can lead to significant increases in the cost of satisfying species requirements at the firms plants. The results obtained for the experiments which incorporated variability in catch rates show the cost increases with increasing variability, but not in a dramatic way. The results obtained from the model can be used to make some generalizations about the industry, particularly if it can be shown that the model can out-perform a companys experience on a given test problem from historical records.<<ETX>>

PLANNING FISH SCOUTING ACTIVITY IN INDUSTRIAL FISHING

Abstract Vertically integrated fish processing firms with a fishing fleet are forever faced with the difficult problem of dispatching fishing effort in such a way so as to minimize the cost of landing plant requirements. This landed cost of fish is a direct function of the catch rates observed on the fishing grounds. Unfortunately, these catch rates are not known a priori. Fish scouting involves routing a trawler over a set of fishing grounds to sample catch rates. The data obtained are used to develop a profile of the expected performance of the fleet. These data are also used to dispatch the rest of the fishing fleet. In dispatching the scouting mission, the dispatcher must select a set of grounds to visit so as the maximize the value of the information obtained while respecting a budget or trip limit constraint. The value of the potential information depends on the species required at the plants, and the age of the current information on the state of catch rates on the various fishing grounds. We present a mathematical programming model to assist the dispatcher in coming up with the optimal mission for the scouting fleet. We demonstrate the model on a problem involving 15 fish stocks.

Planning annual allocation of fisheries surveillance effort

Abstract Fishery surveillance on both of the Atlantic and Pacific coasts of Canada costs the Canadian taxpayer several millions of dollars annually. Few of these dollars are recovered through fines from successfully prosecuted cases involving fishery violations. Consequently, federal fishery surveillance programs are not self-supporting—they are subsidized. In order to reduce the cost of the program to the Canadian taxpayer, it is necessary to implement a cost-effective program which maximizes the deterrent effect of surveillance effort. In this paper we present a tactical linear programming for effectively allocating annual surveillance effort to monitor and control fishing activity, and to enforce fishery regulations in accordance with resource management plans and objectives for the offshore fishery. In particular, we focus on the Atlantic fishery. The tactical model determines on an annual basis, for each of the surveillance units, the amount of effort allocated to each fishery/fishing zone in each time period while maximizing effectiveness. Once the effort allocations are made, operational tasks which involve scheduling surveillance/patrol units on a short-term basis in accordance with the annual plan and current priorities must be completed. We demonstrate the potential use of the planning model.

A two-stage procedure for planning marketing and fishing activities in fish-processing firms

Abstract Planning the activities of fish-processing firms takes place against a backdrop of significant uncertainty. Demand patterns, fish supply, harvesting rates, fish quality, and more, are all uncertain planning parameters. This paper presents a two-stage decision framework for assisting the planning of harvesting and marketing activities in integrated fish-processing firms. These firms operate under an enterprise allocation scheme which gives them quasi-property rights to specified quantities of fish in the sea. The first stage of the decision methodology addresses the aggregate use of the firms resources and enterprise allocations over a 1 year planning horizon. The objective is to maximize potential net profit to the firm. First-stage decisions are subsequently used to set the parameters and boundaries for the second decision stage which deals with the dispatching of the firms fishing fleet. The fleet must be dispatched in a manner which allows the actualization of first-stage decisions at minimum cost. Information obtained from the second stage is used to modify the first-stage decisions if necessary. Using a combination of real and fabricated data, we demonstrate how the framework would be implemented for a typical firm in the Canadian Atlantic groundfish industry.