CA2832806C - Method for optimizing schedule compression in projects using accelerating and overlapping of activities - Google Patents

Abstract

This innovation is in the field of project management and project controls systems for use in compressing projects schedules. Compressing project schedule requires additional cost for more resources, when activities are accelerated, or to cover the resulted rework when activities are overlapped. This causes schedule compression costlier than normal execution. However, there are benefits associated with schedule compression as well. This creates a trade-off between costs and benefits of schedule compression. Performing this trade-off in real industrial projects is very time consuming due to the large number of possible solutions. An innovative algorithm and computer tool is designed and developed to provide optimum degree of accelerating and overlapping of activities in projects. The developed algorithm provides the best or close to the best schedule compression strategy that offers maximum benefit and minimum duration in projects. The algorithm and computer tool are developed using evolutionary optimization techniques to optimize the solutions.

Description

Method for Optimizing Schedule Compression in Projects using Accelerating and Overlapping of Activities Description Field of Invention This invention is in the field of project management and project controls systems. More specifically, the presented method is within the schedule development domain.
Background of the Invention Today's volatile economy encourages project owners to monitor their project schedules very closely, to ensure timely completion. The first incentive for such monitoring is escalation in commodity price and labour costs that may easily become more than the estimated amount. In addition, assumptions about the price and demand for the end-products are based on information available at the time of the initial estimates, which can be years prior to production. Any changes in these assumptions due to market fluctuations may render a project which was previously feasible financially unjustifiable after several years. On the other hand, when end-product prices are much higher than assumptions made during a feasibility study, organizations may be willing to finish projects earlier than planned to take advantage of higher profit margins even with the higher cost. For such reasons, a reliable mechanism for calculating time-cost trade-offs will considerably assist project managers in deciding on the level of schedule compression in projects, to consider reducing the duration with a reasonable cost increase.
Schedule compression is a major step in developing projects' schedules.
Schedule compression is the technique by which project managers shorten project duration. Generally schedule compression is performed using two main techniques: the first is termed "accelerating":
shortening task duration by, essentially, adding more cost. Naturally enough, additional costs may be incurred for extra resources, higher overtime rates or any other payments, for reducing the task duration. The second technique for schedule compression is termed "overlapping".
Overlapping is to perform normally-sequential activities in parallel. Unlike accelerating, activity overlapping often does not add any instant or obvious extra direct cost, but it can potentially add more risks and uncertainties to a project. The sources of these risks mostly derive from starting a successor activity with incomplete data and information. In those circumstances, any changes in the predecessor task's outputs would likely produce rework for the successor task and thus prolong the duration of the task. Hence, overlapping may not be as efficient as expected, due to such uncertainties. Because of the different impacts of these techniques on project time, cost and quality constraints, it is a challenge to develop a reliable analytical tool for performing time-cost trade-offs in cases where both options are available.
There are no known algorithm that compares the costs and benefits of each technique and recommending the best combination of accelerating and overlapping in the schedule compression process.
The purpose of this invention is to develop an algorithm to determine the optimum degree of accelerating and overlapping in combination, to reach the maximum benefit while meeting or advancing the project target dates.
Summary of the Invention This invention determines the optimum degree of accelerating and overlapping among the project activities, to maintain a reasonable trade-off between time saving and the cost of schedule compression. Solely performing activity accelerating adds direct cost to the project.
Too much overlapping means adding the risk of rework to the project resulting from change in the input data and information. This topic is important since complex mega-projects with large number of schedule activities mandate having a reliable algorithm to recommend the extent to which activities should be overlapped and/or accelerated to obtain the maximum time saving with minimum cost increase. Public domain is currently lacking a solid connection between the two major schedule compression techniques.
This algorithm is providing a new methodology to determine the optimum degree of accelerating and overlapping among the project activities by trade-off between time saving and the cost of schedule compression. This algorithm can be used regardless of number of activities.
The introduced algorithm formulates the schedule compression process by connecting the major accelerating and overlapping theories and maintaining a balance between their impacts. It

2 addressees change in schedule critical path and is also capable of functioning in multi-predecessor/successor environment using four types of relationships; Finish-to-Start (FS), Finish-to-Finish (FF), Start-to-Finish (SF) and Start-to-Finish (SS).
The developed algorithm consists of two major sub-programs: Original Population and Optimization. In order to carry out the optimization, the first sub-program creates random solutions as the original population based on the given inputs. The duration of these tasks and the amount of overlapping with each other are all random, and these records will be stored in the database for further optimization. In the next step, the second sub-program will use the original population to optimize the results using cross-over or mutation, and produce new solutions. The cross-over subroutine randomly picks two solutions from the population and shares the characteristics to produce new solutions. The mutation subroutine picks one random solution and changes some of its characteristics randomly. If cross-over or mutation results in better solution, it will be kept and replaces the worst stored in the population; otherwise, it will be ignored and the program will continue producing new solution.
The developed algorithm first runs the accelerating subroutine to come up with the durations and required hours for each task. Then it continues with the overlapping subroutine to determine the overlapping lead times and their rework hours. The overlapping rework hours then will be added to the required hours for the successor task. The calculated hours and cost for each activity are then added to calculate the overall cost of schedule compression, and stored in a separate spreadsheet.
Brief Description of the Drawings Figure 1: Illustrates schedule compression algorithm components Figure 2: Illustrates schedule compression algorithm in 2 parts (Figure 2-1 Preliminary Solutions; Figure 2-1 Optimization algorithm) Figure 3: Illustrates the schematic diagram of the developed algorithm, user inputs, applications' interactions and algorithm outputs

3 Detailed Description of the Preferred Embodiment As briefly described above, there are two main techniques by which schedule compression can be accomplished: activity accelerating, and overlapping.
Activity accelerating is shortening project tasks durations by incurring more cost. Overlapping is to perform the activities in parallel where in normal execution they would be completed in sequence. The amount of overlapping between two activities is expected to shorten duration, though due to the risks of changes and necessary rework, the duration saved would not be exactly the same as the length of the overlap.
This invention provides an optimization algorithm as a tool for determining the optimum degree of accelerating and overlapping in schedule activities. The "optimum degree" is the point that offers the maximum benefit from schedule compression.
Within the industry, the conventional practice for schedule development in projects is to conduct an Interactive Planning Sessions (TAP). This method is a fast approach to developing the schedule, and because it involves all the contributing parties in a session, usually is more efficient. The participants need to address the interdependencies between various disciplines involved in the project, as well as dependencies between the activities.
Using the information obtained during the IAP, planners develop the schedule using the inputs provided in the session. The common observation is that by following the standard sequence and normal duration of activities, it is unlikely to meet the intended major project milestones or completion date. Therefore, the schedule may require compression at the beginning of project or at later dates when delays start to appear. In theory, schedule compression uses two techniques: accelerating and overlapping. In this context accelerating is further sub-divided into crashing and substitution methods as explained more hereafter.
Crashing can be accomplished by increasing the man-hours for faster task performance and earlier completion. Techniques for crashing are overmanning, extended work time and multiple-shift work.

4 Overmanning simply means increasing the number of resources assigned to a task to reduce the duration. Overmanning increases project cost due to inefficiencies involved when assigning more resources to a task, but it is an effective tool for reducing task duration.
Overmanning can be understood in two different ways. First, it can refer to the increase in crew sizes in an amount that exceeds the optimal crew size. The optimal crew size is the minimum amount of workers required to complete a task in the assigned period of time.
Overmanning can also be defined as an increase in the peak number of workers of the same trade over actual average manpower during the project. Both approaches to increasing the number of craftspeople allow progressing at a faster rate, to diminish the time it takes to complete activities.
Overmanning in design projects can also be accomplished by setting up a multiple-office design environment in the project. This means setting up more than one design office to provide engineering and design support services to the project team. These offices can be at the same geographical location or remote offices.
The other crashing method is Extended Work Time or Overtime. Durations of project tasks in a schedule are estimated with the assumption of working for a certain number of hours per day (for example, a typical eight-hour workday). Thus, another method to reduce task duration can be overtime, or extending the number of work hours per day.
Overtime accelerates progress of the activities without the need to engage extra resources, and requires less coordination as compared to overmanning. But overtime usually mandates payments on a higher rate (1.5 or more), after work hours and on the weekends. In addition, when overtime is continued for a long period of time, it will decrease efficiency. This is due to mental and physical fatigue, increased accident rates and reduced safety, increased absenteeism and low morale.
Multiple Work-Shifts is the other alternative for crashing is to set up multiple work-shifts for the tasks. Shift work is the hours worked by a second group of craftsmen whose work on a project is performed after the first or primary work force of the same trade has retired for the day. Like overtime, using multiple work-shifts is an effective way of reducing project duration, as it approximately doubles the amount of work hours per week. Engaging one or two additional shifts can lead to large reductions in activity durations and by crashing the activities on the critical path, overall project duration is decreased.

5 Activity substitution entails the employment of a different technology with a view to achieving the same result. An example would be the replacement of a manual mode of fabrication operation with a computerized numerical control machine. Using a different type or special kind of software to accelerate the design process, engaging other installation machineries or more productive equipment, or sub-contracting a part of the work for faster performance are other examples of substitution.
Activity crashing and overlapping incur both fixed and variable costs;
however, activity substitution involves fixed costs only. Researchers assert that in practice, the major cost of the substitution of an activity, for example employing an alternative technology, resides in the acquisition cost. In oil and gas design projects, substitution can be applied by acquiring and utilization of 3D modeling software such as PDMS or PDS instead of 2D
applications.
Therefore, there is a difference between the calculation of crashing costs and substitution costs.
The crashing cost is the cost of additional resources in terms of hours, while substitution cost would be the cost of a new tool or method.
Due to the similar implications of substitution and crashing in this time-cost analysis, they are classified under one category called "Accelerating,' to simplify the process. The developed model in this patent is focusing on the time-cost trade-off inherent in changing one variable (cost or time) and analyzing the impacts on the other. The decision as to what method to be used to obtain the expected result (crashing or substitution) lies within the project manager's discretion. Generally, activity accelerating practices increase the project cost but rarely add risk to the project. Thus, accelerating is a good solution for conservative decision-makers who are willing to finish the project earlier while taking minimum risk, as they can focus on crashing or substitution of activities on the critical path, as long as the cost of the total project is within the budget.
Overlapping (or "Fast-Tracking") is a schedule compression technique in which phases or activities normally performed in sequence are performed in parallel.
Overlapping increases uncertainties and can result in changes, rework and extra costs as the successor activities must start with incomplete data. Overlapping does not have any visible cost at the beginning but increases the probability of waste and rework in the successor activity as there is a risk of change in predecessor outputs. Therefore, the cost of overlapping would be the cost of resources for

6 extra time needed to complete necessary changes and rework due to overlapping.
A decision about whether a specific degree of overlapping is desirable can be made on the basis of a trade-off between its positive and negative impacts. Generally, it is not recommended to utilize excessive overlapping to compress the project duration, as this makes the project risky due to the potential of cost increases resulting from waste and rework.
Considering above, this invention presents a new methodology to determine the optimum degree of accelerating and overlapping among the project activities, to maintain a reasonable trade-off between time saving and the cost of schedule compression. As discussed earlier, solely performing activity accelerating adds direct cost to the project. Too much overlapping means adding the risk of rework to the project resulting from change in the input data and information.
Today's massive, complex mega-projects with thousands of schedule activities mandate having a reliable algorithm to recommend the extent to which activities should be overlapped and/or accelerated to obtain the maximum time saving with minimum cost increase.
Public domain is currently lacking a solid connection between the two major schedule compression techniques.
The algorithm determines the optimum degree of accelerating and overlapping among the project activities by trade-off between time saving and the cost of schedule compression. This algorithm is suitable for networks with any number of activities. The introduced algorithm formulates the schedule compression process by connecting the major accelerating and overlapping theories and maintaining a balance between their impacts. It is also capable of functioning in multi-predecessor/successor environment using four types of relationships; Finish-to-Start (FS), Finish-to-Finish (FF), Start-to-Finish (SF) and Start-to-Finish (SS).
As explained earlier, the proposed algorithm has two sub-programs: Original Population and Optimization. Figure 1 is illustrating the components of algorithm and their interaction with each other.
1005 is the Original Population sub-program and it is consisting of two subroutines Accelerating (1010) and Overlapping (1015). 1010 first determines the accelerated duration for activities in schedule and then 1015 specifies the overlapping degree between the dependant activities. The degrees of accelerating and overlapping in the original population are all random

7 with equal chance of creation. When sub-program 1 is completed, a pre-defined number of solutions for the schedule are available which will be further used for optimization.
Optimization sub-program (1020) then uses the generated population by 1005 and attempts to improve and produce better solutions. This is carried out by two subroutines of Cross-over (1025) and Mutation (1030). 1025 shares two random solutions' characteristics with the hope of creating a better solution. 1030 is to change one characteristic of an existing solution randomly with the hope of creating a better solution. If the generated solution is better than the worst solution in the population, it replaces that in the population. If the created solution is not better than the old one, program skips that and continues with a new solution.
Figures 2-1 and 2-2 illustrate the details of processes in each sub-program as explained in Figure 1.
First in the algorithm, accelerating subroutine calculates the tasks' durations and hours required randomly (2110). Required hours for each random duration can be read from a database or calculated using a formula. If the selected random duration is normal duration of the task, required hours will be equal to the hours for normal execution. As the task duration is reduced, accelerating techniques (crashing or substitution) come into play and there will be a need for more hours to perform the task in a shorter time. In the next step, overlapping subroutine (2115) calculates overlapping between the activities and rework duration resulted from the overlapping.
Similar to accelerating, rework duration also can be determined using a database or a formula.
The calculated rework duration and pertaining hours are then added to the successor tasks (2115). Then the overall project duration is calculated and resources are leveled (2120). Using the cost of compression and the benefit of saved duration, the net benefit or loss of schedule compression is calculated (2125). If the project duration is acceptable and meets the pre-defined criteria (2130), it will be appended to the solutions database (2135) otherwise voided and algorithm starts with a new solution (2145). This process will be iterated and continued until enough number of solutions is achieved (2140). Required number of solutions depends on the schedule size and number of activities but a minimum of 50 is recommended.
When all the original solutions are created, the optimization algorithm starts improving these randomly generated solutions.

8 First, the algorithm randomly picks two of the generated solutions from the database of original solutions (2210). Attributes of these two randomly selected solutions are being swapped (cross-over) or one attribute of a solution is randomly changed (mutation) to create a new solution (2215). Overall duration of the project is then calculated and resources are leveled (2220). A check needs to be done to see if the calculated duration is acceptable and meets the criteria (2225). If the duration is acceptable, the net benefit/loss of the compression also will be calculated (2230) and if the compression benefit is more than the worst solution in the database (2235), it replaces the worst solution (2240). Otherwise, if the project duration is not acceptable (2225) or the net benefit of the new solution is not better than the worst solution in the database (2235), the new solution is voided and algorithm continues creating a new solution (2250). This process is continued until the intended duration or benefit is reached and at that point algorithm stops running (2255).
The intended solution is the solution that meets both duration and benefit pre-defined criteria. A special feature of this invention is its capability to provide more than one solution and decision maker (project managers) can review various alternatives and pick the best that suits their other technical or safety requirements.

9

Claims (2)

Method for Optimizing Schedule Compression in Projects using Accelerating and Overlapping of Activities Claims

1. A computer implemented method to determine the optimum degree of combining activity accelerating and overlapping using cross-over and mutation of activity characteristics to increase project benefit while compressing schedule, the computer implemented method supports schedule activities with more than one predecessor or more than one successor and further supports four activity relation types of Finish-to-Start, Start-to-Start, Start-to-Finish, and Finish-to-Finish, the computer implemented method produces an original population of initial solutions with random accelerating and overlapping durations, calculating associated rework duration and hours for each degree of overlapping and adding the calculated rework duration and hours to successor tasks in each run, the computer implemented method further comprises checking validity of the project duration for each generated solution, voiding invalid solutions with durations longer than target duration, calculating the schedule compression benefit for valid solutions, and then appending valid solutions to a solutions database.

2. A computer implemented method to determine the optimum degree of combining activity accelerating and overlapping using cross-over and mutation of activity characteristics to increase project benefit while compressing schedule, the computer implemented method comprises randomly picking pair of solutions in each run from solutions database and swapping their attributes which are hours, task duration, accelerating degree, overlapping degree, rework duration/hours using cross-over, or changing one of the attributes using mutation to produce a new resulted solution, the computer implemented method further checks the new resulted solution and if its calculated benefit is higher than one of the existing solutions in solutions database, will replace the lowest benefit solution in the solutions database, otherwise new resulted solution will be voided, the computer implemented method iterates the cross-over and mutation calculations to compress schedule and increase benefit and concludes the calculations either after reaching an intended project duration or reaching an intended project benefit.