B.5.MPC.R - Repair a Physical-Mathematical-Computational Connection

Type: Method pattern Status: Draft Normativity: Normative unless marked informative

B.5.MPC.R:1 - Problem frame

Use this pattern when a physical question has an attempted answer, but the physical account, mathematical reasoning, computation or observed execution no longer fit together. A requested result may have changed, a command may produce an unexpected motion, or an observation may fail to distinguish the physical possibilities that the answer assumes.

First useful move: state the difference that now matters and compare the two contributions at the point where their meanings or results disagree. For a delayed robot stop, compare the duration used in the distance calculation with the duration for which the motor remained active.

The useful result is a repaired connection and its consequence for the physical question, or a located incompatibility that directs the next contribution. The method specializes B.5’s reasoning revision for the joint work described by B.5.MPC. B.5.RR supplies the general treatment of affected reasoning, shared prerequisites and sufficient alternatives.

The practitioner needs the subject knowledge required to interpret the contributions, or access to that help. Use a known adequate correction directly when its conditions fit. Use B.5.MPC when the joint account has yet to be constructed. A fault already located within one discipline can proceed through that discipline’s diagnostic or repair method.

B.5.MPC.R:2 - Problem

Local success can survive a failed joint answer. Algebra can produce a command beyond the device’s range. A frequency calculation can correctly analyze samples while those samples are compatible with several physical frequencies. A voltage measurement can be accurate for the connected circuit while the question concerns the circuit before the meter was connected.

Trying harder within the successful contribution can leave the failure unchanged. More arithmetic precision does not make an unavailable command executable. More samples taken under the same ambiguous sampling relation may preserve the ambiguity.

The difficulty is to locate which connection fails and change the contribution that can repair it. Several accounts may explain the same discrepancy, so an observed disagreement must not be assigned to a particular component before the available reasoning distinguishes those possibilities.

B.5.MPC.R:3 - Forces

B.5.MPC.R:4 - Solution

Keep the receiving physical question in view while tracing the failed connection:

State the changed result or disagreement → recover the contributing relations → locate the incompatible use → construct a sufficient repair → carry it through the affected contributions → use the revised physical consequence.

These are dependencies of the repair. A known physical change, a failed calculation or a surprising observation can each be the starting point.

B.5.MPC.R:4.1 - Compare results for the same physical question

Name the physical difference that the answer must resolve and the conditions under which it is wanted. Put the attempted result beside the changed request or observation. Recover the participants, units, times and operating regime needed to compare them.

For instance, a predicted displacement during eight seconds of motor operation and a measured displacement during eleven seconds concern different durations. Aligning the comparison may already explain the discrepancy.

If the question changed, say which quantities are now supplied and which must be determined. Keep the physical relation available while choosing its new use. A velocity-to-duration calculation can become a duration-to-command calculation.

If the comparison concerns a performance claim, recover the observation and its measurement relation. An acknowledgement can supply information about command reception while leaving physical completion unobserved.

B.5.MPC.R:4.2 - Recover what each contribution supplies to the next

Start at the mismatch and follow the needed relations in both directions. Ask what the receiving operation takes as input, what the supplied result actually represents and which condition permits that use.

The following questions distinguish common repair locations. Use the ones that can explain the disagreement.

Explain an unresolved connection in the working notation. Use B.5.RA to recover a needed argument and B.5.RC to recover a construction. C.29.1 compares mathematical transfers; C.29.2 constructs a computation; C.29.3 compares a computation with its realization. Their results answer the corresponding rows, while the physical question determines which row matters.

Keep a common premise visible across different contributions. An analytical calculation and a simulation that both assume no slip leave the same question open when slipping is suspected. A different algorithm is a sufficient alternative only if its premises and output interpretation fit the receiving use.

B.5.MPC.R:4.3 - Locate the incompatibility without guessing its cause

Work a small instance or derive a consequence that distinguishes the competing accounts. If a command is outside the admitted range, its value and range already establish that incompatibility. If two physical inputs produce the same represented data but require different answers, exhibit those two inputs and their common representation.

When several explanations remain, compare what each would require you to change. A motion discrepancy might come from calibration, a retained command or an incorrectly interpreted position. Recover an available observation that distinguishes those accounts, or choose a worthwhile further observation through C.11.DUA. An investigation is unnecessary when a sufficient bound or conditional answer already settles the present decision.

Account for what the observation does. Identify which processes continue during measurement or debugging, which input remains applied and which state is changed by the probe. Use A.3.3 for the state and C.16 for the measurement relation when that construction is unresolved.

A mathematical account may support several physical interpretations. Compare their additional physical assumptions and the observations or interventions that distinguish them. Retain an unresolved difference when the available observations do not decide it; use the consequences shared by those interpretations when they suffice.

B.5.MPC.R:4.4 - Construct the repair that the receiving use needs

Choose the contribution that can remove the located incompatibility. The choice follows its cause and the wanted result.

• If the requested unknown changed, retain the interpreted relations, release the former fixed value where appropriate and derive the new computational direction.
• If a representation merged cases that the question must distinguish, retain the missing information, obtain a different observation or use a result that the merged representation still supports.
• If the procedure uses the wrong operation, constraint or approximation, replace that part and establish the property the receiving use needs.
• If the physical or measurement account omits a consequential interaction, model that interaction or change the arrangement so that the retained account applies.
• If execution differs in preparation, range, state or completion, revise that part of the realization or choose an available realization that supplies the required behavior.

An expression of an equation states a relation; an assignment or solver performs a chosen operation. Keep the equation’s meanings when generating or editing that operation. A modelling environment can help transform the equation system, but the physical interpretation and the requested result still guide the choice.

Compare the repair with a sufficient alternative. Changing a deadline may be cheaper than replacing an actuator. A useful ambiguity result may suffice before changing measurement equipment. Present a changed requirement as a choice for the receiving work.

B.5.MPC.R:4.5 - Carry the repair through its dependents

Use B.5.RR to derive the affected consequences. Reuse contributions whose conditions and meanings still fit; revise each needed use of a changed shared premise.

Compare the repaired contributions on the same input or physical case. For a mathematical change, derive the needed property. For numerical work, carry the error relevant to the physical question. For an execution claim, compare the prepared state, operation and interpreted observation with the intended result.

Keep the comparison at the claim’s scope. A conditional calculation can establish what a proposed arrangement would do under its assumptions. Actual performance requires the observations and inference appropriate to that use.

If the comparison still fails, follow the remaining disagreement. Do not repeat a test that leaves the live alternatives indistinguishable. Change the question or obtain the missing contribution when the present means cannot settle it.

B.5.MPC.R:4.6 - Return the physical consequence and the next useful action

State what can now be understood, constructed or changed. Return a usable command, corrected interpretation, realizability bound or a remaining ambiguity with its effect on the receiving decision.

When work is divided, give the next contributor the quantities and conditions at the unresolved connection, the result needed and how it will be used. The receiver must be able to reconnect the supplied contribution to the physical question. A.15.9 helps obtain that contribution.

The repaired reasoning may reveal another worthwhile problem: extending an operating regime, constructing a more informative measurement or developing an operation for a changed class of questions. Carry that possibility into B.5’s next inquiry when pursuing it would enable further work.

B.5.MPC.R:5 - Archetypal Grounding

These are constructed cases under stated models. They demonstrate different repair locations and the contributions that follow from them.

B.5.MPC.R:5.1 - A robot receives a new deadline, then keeps moving during a pause

A robot must travel d=1 m along a straight guide. In the stipulated regime, constant speed obeys v=k*u, where u is a dimensionless command and k=0.2 m/s. The model initially idealizes starting and stopping as immediate. The available command range is 0<u<=0.8.

With u=0.5, the relation d=vT gives T=d/(ku)=10 s. The question then changes: find the command for T=5 s. Keep the relations but choose u as the unknown:

u=d/(kT)=1/(0.25)=1.

The algebra supplies a value outside the available range. The corresponding lower bound on duration is T>=1/(0.2*0.8)=6.25 s. This answers the feasibility question before another controller is written. Changing the deadline or the available motion capability is now a meaningful choice.

Suppose the requester chooses T=8 s. The required command is u=0.625, giving v=0.125 m/s. A controller applies that command, counts eight seconds of active program execution and then issues stop.

During a diagnostic run, a debugger freezes the program and its timer for three seconds. In this example the output retains u=0.625 and the motor continues to run. The stop is therefore issued after eleven seconds of physical operation, giving d=0.125*11=1.375 m under the model.

The disagreement concerns execution timing. The equation and chosen command remain appropriate to eight seconds of motion. If the question concerns an uninterrupted run, observe that run under its intended timing. If such pauses can occur in the required operation, provide a stopping mechanism whose elapsed-time behavior survives the pause, or revise the control arrangement accordingly. Test its physical completion at the condition the use needs.

For example, suppose an independent timer can remove the motor command and continues to run while the controller is paused. Start it with the command and set it to remove the command after eight physical seconds. Under the example’s immediate-start/stop model, motion then lasts eight seconds and covers 0.125*8=1 m, including a run with the three-second program pause. Compare the timer’s actual behavior with those conditions before relying on that performance. If no such mechanism is available, the next contribution is a way to stop motion with that timing or a revised operating requirement.

A real stopping delay or speed transient would be another changed premise. Incorporate it when predicting final displacement; acknowledging stop does not measure that displacement. The example’s next useful result is a duration bound or a repaired timing design with an identified performance question.

B.5.MPC.R:5.2 - A correct spectrum calculation answers an ambiguous physical question

A measurement uses uniformly spaced samples at f_s=100 samples/s. The earlier physical account admits a single sinusoidal signal with frequency below 50 Hz. A spectrum calculation returns a 10 Hz component, interpreted under that restriction.

The apparatus is changed and the admitted frequency range becomes 0 to 120 Hz. The same interpretation is now in question. For ideal samples of a unit-amplitude cosine at times n/100 seconds:

cos(2π90n/100)=cos(2π10n/100)

for every integer n. This follows because 90/100=1−10/100 and cosine is periodic and even. A 110 Hz cosine gives the same samples as well.

Consequently, the sampled sequence fits physical frequencies 10, 90 and 110 Hz in the new range. The spectrum calculation can remain correct for its input. Increasing its arithmetic precision or merely collecting more samples at the same times leaves this ambiguity.

To determine which frequency is present, change a contribution that distinguishes those cases. For the stipulated single-tone model, sampling at 300 samples/s places the whole admitted range below half the sampling rate. A three-second record of a 90 Hz signal then contains 270 cycles; the spectrum can return 90 Hz under the stated ideal timing and signal assumptions.

For a physical measurement, obtain the needed sampling and input-conditioning behavior from the instrument account. Frequencies outside the admitted band, timing error and an unsuitable analog input path can reopen the interpretation. A filter that removes the signal of interest changes the measurement question rather than recovering its frequency.

The repair is a changed acquisition and interpretation, with a recomputation on the new data. The receiving user can now distinguish a physical-frequency estimate from an unresolved alias. If the old record is all that is available, return the remaining alternatives instead of selecting one without further grounds.

B.5.MPC.R:5.3 - The voltmeter changes the circuit being inferred

An ideal 10 V source drives two 1 MΩ resistors in series. The wanted quantity is the midpoint voltage before a meter is connected. The divider relation gives 5 V.

A voltmeter with 1 MΩ input resistance is connected between the midpoint and the lower terminal. In the connected circuit it lies in parallel with the lower resistor, whose combined resistance becomes 0.5 MΩ. The same divider construction now gives:

V_mid=10*0.5/(1+0.5)=10/3 V.

The indicated value can agree with the connected-circuit model. Comparing it directly with the unloaded prediction had omitted the meter’s interaction.

One repair is inferential: with these resistor, source and meter models, recover the unloaded value from the model, keeping its conditional status. Another is experimental: choose a measurement arrangement whose loading is small enough for the wanted use and account for its remaining effect. Repeating the same connection alone preserves the load.

The result tells the practitioner which voltage was observed and how the wanted voltage can be obtained. A different observed value may require revisiting the assumed source, resistor or instrument behavior through their subject methods.

B.5.MPC.R:5.4 - A motion command changes its unit and starting point

In C.29.3’s robot case, the effective wheel radius is 0.05 m, ten motor revolutions turn the wheel once, and the interface counts 1,000 increments per motor revolution. A two-metre displacement requires a total of 63,662 motor increments after rounding. C.29.3 handles sending that total through the available command field.

Now change the radius to 0.06 m and the interface to an absolute wheel-position target. It counts 1,000 units per wheel revolution, currently reads 3,000 and admits targets from 0 to 65,535. The requested forward displacement remains 2 m. Assume rolling without slip and completion at the commanded count.

Recover the count’s new meaning before reusing the calculation. If q is the target, the wheel travels (q−3,000)/1,000 revolutions, so the model gives:

d=(q−3,000)/1,000 * 2π * 0.06 m.

The new scale already counts wheel revolutions; use that scale to obtain the needed increment:

Δq=2/(2π*0.06) * 1,000 ≈ 5,305.16477.

Round to 5,305 and add the initial count, giving target q=8,305. This target is admitted. Its modeled displacement is approximately 1.99993788 m. Nearest-integer rounding contributes at most half a count, corresponding to π*0.06/1,000 m ≈ 0.00018850 m; compare that contribution with the receiving accuracy requirement.

Reusing the former total 63,662 as the absolute target would pass the new range check but give approximately 22.8690 m under this model. The value has changed from a relative motor increment to an absolute wheel position. The repair therefore changes the physical parameter, the count-to-motion correspondence and the computational use of initial state together.

The useful return is the new command with its displacement and rounding consequence. A changed starting count requires recomputing the target; a changed radius or count scale reopens their correspondence. A slip or incomplete motion requires the relevant physical or execution repair. C.29.3’s preparation and completion comparison can then use this revised command.

B.5.MPC.R:6 - Bias-Annotation

The cases use elementary models so the connection can be worked explicitly. Nonlinear dynamics, statistical observations or coupled physical computations can require more demanding subject contributions. Preserve their interaction and uncertainty conditions when applying the same repair method.

A visible software failure can attract attention even when the computation is correct and its physical interpretation is wrong. Compare the meaning of the input and output before choosing a programming repair.

B.5.MPC.R:7 - Conformance Checklist

• The attempted and wanted results concern a stated physical question and comparable conditions.
• The contribution supplied at the disagreement has an interpretable meaning and receiving operation.
• The repair distinguishes a located incompatibility from an unresolved diagnosis.
• Shared premises remain visible across calculations, simulations and observations.
• The repair changes the responsible relation, procedure, realization or requested result.
• Its dependent consequences are reworked and returned at the supported scope.
• The receiver can perform the next action or identify the contribution still needed.

B.5.MPC.R:8 - Common Anti-Patterns and How to Avoid Them

B.5.MPC.R:9 - Consequences

The method makes a failed joint answer useful: the disagreement can reveal a range limit, a lost distinction, an omitted interaction or a missing construction. It also supports division of work by the contribution needed at the failed connection.

The cost is reconstruction across disciplinary boundaries. A sufficient conditional result or a known correction can keep that effort small. Where several causes remain compatible with the evidence, the result can remain a bounded diagnostic question.

B.5.MPC.R:10 - Architectural Rationale

The common revision method B.5.RR follows affected reasoning. This specialization adds the recurring correspondences that a physical application consumes: a physical account expressed mathematically, an interpreted computation and a realized or measured operation. It lets the practitioner choose the repair that can change the physical answer.

Working from the disagreement can settle the question with a bound or identify the observation that must change.

The physical and mathematical accounts need each other without becoming the same account. A mathematical equality can expose why two frequencies have identical samples. The physical sampling arrangement determines whether those are the relevant possible inputs. Computational analysis then operates on the resulting data. The useful knowledge comes from keeping those relations together; applying it can change the measurement method itself.

B.5.MPC.R:11 - SoTA-Echoing

How can an expression help locate a failed connection? Adapt Sussman and Wisdom’s Structure and Interpretation of Classical Mechanics, second-edition preface (2015): expressing mathematics as executable procedures exposes operations that informal notation can leave implicit. Sections 4.2 and 4.4 use that feedback where available. Compared with polishing a symbolic answer alone, it can locate an unusable operation; its extra formalization is worthwhile only for the receiving question.

What survives a changed computational direction? Adapt the separation of equation models and generated computations in ModelingToolkit’s version-10 documentation, alongside Ma et al.’s 2021 account. Section 4.4 retains relations while changing givens, unknowns or the obtaining procedure. This supports acausal modelling rather than making one assignment direction part of a physical law. An adequate explicit formula remains cheaper for the robot’s small calculation; no modelling platform is required.

When must observation be treated as an operation? Adapt Khrennikov’s Contextual Measurement Model (2024), §§2.1–2.2: the measurement procedure can change the context used by subsequent inferences. Section 4.3 therefore recovers that operation when it matters. The passive-observation account remains useful where its approximation is adequate. The voltmeter case supplies an elementary circuit construction of an interacting probe; broader physical interpretations need their own assumptions and observable consequences.

Which sampling repair changes the physical inference? Use NI’s sampling and aliasing account, “Sample Rate” and “Aliasing” for the acquisition constraint and input conditioning. Section 5.2 derives its alias equality explicitly and states f_s and the admitted signal band. Compared with increasing numerical precision, changing acquisition can distinguish the physical cases. The instrument’s bandwidth, filtering and timing remain part of that repair.

Reopen the source choices when a physical theory, representation, computational method or executing technology supplies a better way to resolve the same disagreement. Compare the actual operation, consequence and conditions that let the receiving work use it.

B.5.MPC.R:12 - Relations

• B.5.MPC constructs the joint physical, mathematical and computational answer; this method repairs an attempted connection.
• B.5.RR revises affected reasoning. B.5.RA and B.5.RC recover an argument or construction needed for that revision.
• C.29.1, C.29.2 and C.29.3 supply mathematical transfer, computational formulation and realization.
• A.3.3 constructs the relevant state and continuation account. C.16 constructs and qualifies the measurement relation.
• A.6.3.RT helps express the changed relation or operation. A.15.9 helps obtain another practice’s contribution.
• C.11.DUA compares further effort with its effect on the receiving decision. B.3 and B.3.3 supply claim-specific assurance when required.

B.5.MPC.R:End