(1) Explain your problem, don't simply post "This isn't working". What were you doing when you faced the problem? What have you tried to resolve - did you look for a solution using "Search" ? Has it happened just once or several times?
(2) It's also good to get feedback when a solution is found, return to the original post to explain how it was resolved so that more people can also use the results.
Hello,
I would like to get some advice on the following scenario and solution. Any better ideea are much appreciated.
Input Data:
00. 4 machines will be deliverd on sites, 2 big, 2 small. It is necessary to use 1 resource for each equipment - this machine 1, machine 2, machine 3, machine 4;
01. machines will be delivered on 4 sites, on different dates;
02. spacial constraint: maximum machines per site is 2;
03. the movement from one site to another is 3 days;
04. each machine has different productivity, big machine 100 sqm/h, small machine 80 sqm/h.
Required solution:
Optimize schedule in such a way that the activities are executed continously (or with minimum float) between them and find the order of movement of the equipment from one site to another.
My approach:
01. 4 milestones for resource production. Each milestone will produce 1 resource;
02. Created one resource skill with all 4 machines;
03. Activities have assigned 2 teams (maximum equipment per site), each team having the previous created skill;
04. The resource skill is of Quantity type, activities are of productivity type, resources in skill have productivities assigned.
05. Calculation and adjustment:
05.01 Step 1: scheduled with resource constraints to find activity durations and machine allocation for each activity;
05.02 Step 2: introduced activities for machine movement and linked them to the execution activities;
05.03 Step 3: introduced negative resource production at the end of each execution activity for machines that are used on the activity, and introduces positive resource production for machines that are used for the execution activity at the end of the movement activity. This way the the next activity will be delayed by the movement activity.
Is there any better solution?
Best regards,
Bogdan
Printer-friendly version
Send to friend
Rafael,
thank you for your advice and point of view on the matter. Both are highly appreciated.
Best regards,
Bogdan
Thank you for your input on this use-case. Your point of view on this model has been insightful.
Best regards,
Bogdan
Bogdan,
I cannot see units of volume. I cannot see what the machines do. If I delay start to 11/18/24 08:00 of your model duration is reduced to from 345 to 308 days. If I let the software to select machine production project duration is reduced. Project scheduling requires full understanding of all constraints. I do not understand too many variables.
I do not like the idea of making manual adjustments to resource assignments, this can be a source of errors if some activities are delayed and activity sequence changes. A tiny (if compared to total project duration) schedule variance of an activity can become a cause for substantial rearrangements in the project network, it can change critical path or order of activities, enforce a total reorganization of works. Better if the software takes care of it.
I do not like the idea of scheduling the machines on independent teams. Better if the crew is the same from beginning to end. No problem with different crews if each crew work on different shifts.
I am retired and my time is not an issue but the lack of understanding of the job. I am opting out of this discussion.
Best Regards,
Rafael
Bogdan,
I just downloaded the files, I will take a look in detail latter.
Best Regards,
Rafael
Rafael,
Here are the files:
Here are version 50, 60, 500:
https://drive.google.com/drive/folders/1HXcFHdFSQ8i4zw1BcvgFSxiNJg6rOodn?usp=drive_link
Here are my files:
https://drive.google.com/drive/folders/14MoETNIzN4YQ-LR9MXW4TyOiyuCm9iOd?usp=drive_link
One note:
I added Version 04, in which I added predecessors to the working activities. The only thing I had to change is to create a link between the predecesor of first execution activity because Machine 1 arrives after predecessor and leaved before execution activties, since it is not used at all (which makes scense).
Best regards,
Bogdan
Can you tell me which version of Spider you used to create the model?
Spider Project version 24.09.032 - of 17.07.2024 a version older
I just downloaded the new one. With the new version I am still having crashes or odd things.
I believe I could not download all the files as I could only download from one of the two references. I could not download your model versions.
From your image, while the activity descriptions have no meaning to me and I cannot see predecessors to productivity activities it looks good to me. I suspect that if you add predecessors it might or might not create additional idle time, same as in my sample schedule.
Best Regards,
Rafael
PS: I use free Google Drive 15GB no limit on downloads, after many years I am using less than 10% of free limit.
Rafael,
we are on the same page, but we try different approaches. I tried to recreate your model, without looking at the file and I could not do it. I Downloaded the file, looked over it and changed to my file 2 things:
01. Added resource skill to machine delivery;
02. I cheked the assigment priority for resource skills as you, "productivity". The rest is the same.
I did not have the same result. Then I took your model and performed a simple scheduling (CPM), then resource constrained scheduling and it works.
Odd things do happen. If I do not make changes to your schedule everything works as you describe, however if I REPLACE skill assignments, re-enter produtivity or if I delete skills and allocate them again, same number, same produtivity (Machine 1 100; Machine 2 80; Machine 3 100; Machine 4 80) Spider crashes (Version 50 and 60).
Can you tell me which version of Spider you used to create the model? Mine: 24.09.048
https://file.io/brD8T7ZvyJWZ - files are deleted after download
https://file.io/brD8T7ZvyJWZ - files are deleted after download
One more thing. If I combine your original file, with my ideea, that machines arrive in fixed points in time, and the schedule has to adapt to when they arrive (my original problem). Then I get the following result, which is not good.
(Version 500)
The file of this schedule is in the above link.
My main problem, is not to calculate when machines should arive, but how to obtain the best schedule according to when they arrive, and how to move them 1 by 1 or 2 by 2 (when optimal) in such a way that they work constantly and obtain the shortest schedule.
Here is my proposal.
Milestones for Machine Deliver, Milestones for Machine Removal (Link FS +1 Day). It is safe to say that Machines will work at least 1 day.
You schedule the file however you like but not with Fix skills because it cancels Team 1. Then allocate to the removal activities the scheduled resources from the team, and rechedule. Skills resources will be fixed and you will be able to calculate duration of project with relocation, but also the moment of removal.
First Step: Priority on activities (last activity in chain - the removal activity, on productivity of resources in skills, standard (since optimization does not yeld a shorter schedule).
Second Step: Assign to removal activities resources from skill assignments, and the reschedule. The result.
The files are here:
https://file.io/elSlKCLoz0kp - files are deleted after download
https://file.io/elSlKCLoz0kp - files are deleted after download
Best regards,
Bogdan
Vladimir,
I created a long duration activity to see idle resource diagram beyond machines usage as shown in the first image. I am having issues with adding another resource deletion milestone similar to RETURN M3 OR M4, as soon as I create the milestone to RETURN M1 OR M2 and run resource leveling the schedule crashes.
What am I doing wrong? I suspect it might be an issue with how I am using skills.
Bogdan,
Resource Gantt can disclose order of equipment movement as shown in the second image. I just added a user defined field for the area and with a simple formula using the first character of the activity codes it can get populated. If such coding is unavailable Populate downward by WBS level can be of help.
While it looks we are not in the same page of what is exactly to be modeled I hope this can be of some help.
Best Regards,
Rafael
By creating the logic for the software to create and return the machines I got a schedule with no machine idle except during no work days in this case the weekends.
Best Regards,
Rafael
https://drive.google.com/file/d/1Prin71qUzfJIYR1b26cpLKEblE3l7ok1/view?u...
Vladimir,
I liked your Stacked columns diagrams. I used them in my schedule which is for a different model and came out as follows.
Bogdan, thank you for an interesting task.
I tried different approach that also have some problems that must be solved.
In Spider Project resource assignments can be delayed from activity start, but activity can start only if some resource starts its execution. So I created a dummy resource that is produced together with other resources and is assigned together with other resources in separate teams.
Original resources are assigned with 10 days delays (time required for resource delivery).
Dummy resource has small productivity (100 times lower than productivity of original resource) but it still create inaccuracy.
Hello Rafael,
Required solution:
Optimize schedule in such a way that the activities are executed continuously (or with minimum float) between them and find the order of movement of the equipment from one site to another.
I should have been more clear, activities that use the machines, must be performed continiously. Activities can have maximum 2 resources, the ideea is to find the solution for minimum duration while machines must be used constantly, without idle time, or with minimum idle time; relocations are idle time but necessary for correct modeling. In yor model you have delivery, but no relocation.
Each site starts with 0 resources, and the milestones in my model represent time of delivery on site. The scheduling of activities can start the moment the firs machine gets on one site; predecessors and succesors are more flexible from resource point of view. If I use only one team in my model, as you did in yours I get the following result.
By using a single team, and optimization+ algoritms, and priorities, not fixing skills (fixing skills crashes Spider), the result is an error, saying I need more resources:
The only solution, I found, for activities to start work imediatley after the machines arrive is to use Teams, each machine can work independently. After you calculate machine delivery and work, then you must fix resource skills and create activities for relocation and then reschedule to see the final duration.
This is step 1:
This is step 2:
If you do not use strict links in this model, you get:
If you use strict links in this model, you get teams that are no longe scheduled:
If you use negative resource production at the end of the activities, and prositive at the end of the relocation activities, you get the right result (but NO strict links):
(I made relocation activities 10 days so it is easier too see resource realocation).
If you use strict links, you will not get right result, somehow they interfere with resource production:
I hope this time I have explained myself better.
Best regards,
Bogdan
Required solution:
Optimize schedule in such a way that the activities are executed continuously (or with minimum float) between them and find the order of movement of the equipment from one site to another.
I understood that the Deliveries as well as their predecessor activities (necessary for work execution) and successor activities (the work execution activities) must work continuously. In my model each Pase represents all these activities working on the same area using same machine. That deliveries to each area are to be scheduled just in time. My model does that and it works.
Now you are showing some images that spell a thousand words which is good, very good. Still I cannot see the work execution activities (the production activities that make use of the resource) in your model. Because the resource productivities are different it is essential for the production activities be of the production type instead of duration type. On the other hand, deliveries are of fixed duration. Predecessor activities can be many, each of different type, for simplicity I used a single predecessor activity to each production activity.
What am I missing?
Best Regards,
Rafael
Rafael,
you could use maximum 2 resources/activity by adding 2 teams, each machine works independently.
See below picture.
If I check the fix resources, for some reason (I belive this is not right) Spider cancels Team 1. I do not know why. So I will fix resources manually.
However in my example if I make one link strict for example for Site 1 to Relocation then for some reason Spider will assign that sequence the highest priority and activity priorities will not be taken into consideration:
Without strict link:
With strict link:
If I try phase priorities the result is the same (this is the Standard Algoritm and should take activity priority into consideration).
In the end if you use resource allocation for relocation activities or resource production (negative at the end of the Site 1 activity, for example, and positive at the end of the relocation activity) you must still fix resources otherwise you may end up with another combination when recheduling.
This is a bit of a back and forth procedure and with more complex scenarios, like 20+ activities, it takes some time.
I was hoping that Spider would somehow fix resource skill assignment according to the availability of resources based on negative and positive production. I do not even know how to phrase it: Spider should fix resources for each team, then check if the activity has at the end negative production of the same resoruces that are being used on the execution activity and if so, assign highest priority to the next positive production activity that has no constraint (start NET, finish NLT) and reschedule the remaining activities based on the first production activity.
Maybe Vladimir can give us a better ideea on how to model this. Also it would be nice if the user could see resoruce production in the Resource Gantt.
Best regards,
Bogdan
Bogdan,
I revised the model to consider productivities. The baseline use strict links but yields an unfeasible schedule. Version 2 links delivery activities to other deliveries activities. It is feasible but takes one week longer.
A better way to make it feasible is with the use of a Delivery Resource (Quantity 1) as we are not creating a sequential chain.
I find using one named resource for every machine gives me more control. Say M1 M2 .... Then by showing iddle assignments I can delete specific machies from Skill individual activity assignments. I can also change individual assigment productivities. Also convenient for modeling automatic seasonal productivities.
Good Luck
Bogdan,
You are correct. I am still working on 02. constraint: maximum machines per site is 2.
Please note my prior schedule does not consider productivity as all activities are of duration type, no matter the selected resource duration will not change, my mistake.
Best Regards,
Rafael
Thank you for posting your example. I have something similar, but machines are produced (delivered) at 1 or 2 month distance between them.
However, if the location requires 2 or more machines, then they need to be setup in different teams, with quantity of 1 for each skill. You cannot allocate to a single Team quantity 2 or more for the team because it will be delayed untill all resources are produced.
It should do it. Fix resource skills might help to keep machine selection as per selected WBS.
Hoover image>Click>Zoom
Email: PPadmin@planningplanet.com
Replies